Intelligent torque control method suitable for high-torque drilling screw

Abstract

The invention relates to the technical field of drilling engineering, and particularly discloses an intelligent torque control method suitable for a high-torque drilling screw rod, and the method comprises the steps: collecting drill string vibration frequency data and bit pressure adjustment frequency data in real time in a drilling process, extracting a zero-sequence vibration frequency characteristic value in combination with sparse Fourier transform, and evaluating the vibration stability of a drill string; meanwhile, differential adjustment frequency is analyzed through empirical mode decomposition and Hilbert transform, differential adjustment frequency characteristic values are constructed to measure bit pressure adjustment stability, the two characteristic values are fused into a comprehensive characteristic vector, the comprehensive characteristic vector is input into a risk prediction model trained based on a random forest algorithm, and the bit pressure adjustment stability is improved. Intelligent identification of whether the drill column is in a high-frequency resonance critical state or not is achieved, and when it is judged that resonance risks exist, the system automatically triggers a torque compensation mechanism, and bit pressure and torque output are dynamically adjusted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of drilling engineering, and particularly relates to an intelligent torque control method suitable for a high-torque drilling screw rod. BACKGROUND

[0002] In the modern exploration and development process of oil, natural gas and unconventional energy, drilling operations are facing challenges such as complex formation, increased well depth, and increased demand for drilling pressure and torque, especially in high-torque drilling conditions. The drill string system bears a large torsional stress and periodic vibration. As the core power device in directional drilling, the screw rod is prone to high-frequency resonance due to factors such as geological mutations, improper drilling pressure adjustment, or coupling of excitation frequency and natural frequency, which causes severe torsional vibration of the drill string, and further causes drill fatigue fracture, connection thread failure, downhole tool damage, and even wellbore instability, seriously affecting drilling efficiency and construction safety.

[0003] The prior art has the following disadvantages: In the process of intelligent torque control system in high-torque drilling, if the natural frequency of the drill string and the formation or downhole tool is coupled, high-frequency resonance may occur. When the system adjusts the torque in real time, if the algorithm response is delayed or the sensor data has noise, the adjustment frequency of the motor power or the drilling pressure may coincide with the natural vibration frequency of the drill string, forming a positive feedback loop. At this time, the drill string will accumulate severe torsional vibration energy in a short time, and the local stress will far exceed the material fatigue limit. Especially in the weak parts of the drill string joint and thread, instantaneous torsional load concentration may occur, which may eventually lead to drill string fracture. If the resonance is not inhibited in time, it may also trigger a chain reaction, such as drill falling off, wellbore instability, and even downhole tool damage.

[0004] Therefore, it is urgent to propose an intelligent torque control method suitable for high-torque drilling screw rod, which can fuse multi-source sensing data, realize real-time evaluation and intelligent prediction of the vibration state of the drill string and the stability of the drilling pressure adjustment through advanced signal processing and machine learning technology, and actively intervene in the control parameters before the critical state of high-frequency resonance occurs, so as to improve the safety, stability and intelligent level of drilling operations. SUMMARY

[0005] The present application aims to provide an intelligent torque control method suitable for high-torque drilling screw rod to solve the problems in the above background.

[0006] The purpose of the present application can be achieved by the following technical solutions: An intelligent torque control method suitable for high-torque drilling screw rod, comprising the following steps: S1: collecting vibration frequency data and drilling pressure adjustment frequency data of the drill string in real time during drilling; S2: Extract the zero-sequence vibration frequency component of the drill string from the vibration frequency data. Calculate the zero-sequence vibration frequency characteristic value based on the change rate of the zero-sequence vibration frequency component to evaluate the vibration stability of the drill string. S3: Calculate the differential adjustment frequency based on the WOB adjustment frequency data, and calculate the differential adjustment frequency characteristic value according to the degree of change of the differential adjustment frequency to evaluate the stability of the WOB adjustment; S4: constructing a comprehensive feature vector from the zero-sequence vibration frequency eigenvalue and the differential adjustment frequency eigenvalue, analyzing the comprehensive feature vector using a machine learning model, and determining whether the drill string is in a high-frequency resonance critical state based on the analysis results; S5: If the drill string is determined to be in a critical state of high-frequency resonance, the torque compensation mechanism is automatically triggered to dynamically adjust the drilling pressure to ensure safe and stable operation of the drill string.

[0007] Preferably, the process of obtaining the zero-sequence vibration frequency characteristic value is: During the drilling process, the vibration frequency data of the drill string is obtained, the zero-sequence vibration frequency component of the drill string is extracted from the vibration frequency data, and the zero-sequence vibration frequency characteristic value is calculated according to the change rate of the zero-sequence vibration frequency component. Specifically, the following steps are performed: Apply sparse Fourier transform to the time series of the drill string vibration frequency data to obtain its sparse representation in the frequency domain: ; Where, represents the sparse spectrum result, x ( n ) represents the time series of vibration frequency data, n Indicates the number of collection points, k is the set sparsity, indicating the number of main frequency components, represents the frequency component; Identifying zero-sequence vibration frequency components from sparse spectrum , and calculate the rate of change of the corresponding frequency component over time: ; Where, Represents the zero-sequence vibration frequency component, t j Indicates the j Collection time points, j Indicates the number of acquisition time points, Indicates the j The zero-sequence vibration frequency value extracted at each acquisition time point, Indicates the j The rate of change of the zero-sequence vibration frequency value at each acquisition time point; According to the rate of change of the zero-sequence vibration frequency value, the zero-sequence vibration frequency characteristic value is calculated, and the calculation expression is: ; Where,F zv represents a zero-order vibration frequency characteristic value, m represents the total number of acquisition time points.

[0008] Preferably, the evaluation of the vibration stability of the drill string specifically comprises: During the drilling process, it is determined whether the zero-order vibration frequency characteristic value is greater than or equal to a preset threshold value, if yes, the vibration of the drill string is unstable, and if no, the vibration of the drill string is stable.

[0009] Preferably, the process for obtaining the differential adjustment frequency characteristic value is: During the drilling process, the weight-on-bit adjustment frequency data is obtained, the differential adjustment frequency is calculated based on the weight-on-bit adjustment frequency data, and the differential adjustment frequency characteristic value is calculated according to the degree of change of the differential adjustment frequency, specifically comprising: The differential adjustment frequency of the time series of the weight-on-bit adjustment frequency data is calculated, specifically the difference between adjacent time points is calculated to obtain the differential adjustment frequency, all differential adjustment frequencies are integrated into a differential adjustment frequency time series, the differential adjustment frequency time series is subjected to empirical mode decomposition to obtain a group of intrinsic mode functions, and further Hilbert transform is performed on each mode function to obtain instantaneous differential adjustment frequency, and the calculation expression is: ; In the formula, t represents a time series acquisition point, ω i (t) represents the instantaneous differential adjustment frequency of the i th intrinsic mode function obtained after Hilbert transform, represents Hilbert transform, a represents an imaginary unit, i represents the number of intrinsic mode functions, and arg represents the amplitude angle of a complex number, represents a derivative operation, and IMF represents an intrinsic mode function. The standard deviation of the instantaneous differential adjustment frequency of all intrinsic mode functions is calculated to obtain the differential adjustment frequency characteristic value.

[0010] Preferably, the evaluation of the stability of the weight-on-bit adjustment specifically comprises: During the drilling process, it is determined whether the differential adjustment frequency characteristic value is greater than or equal to a preset threshold value, if yes, the weight-on-bit adjustment is unstable, and if no, the weight-on-bit adjustment is stable.

[0011] Preferably, the analysis of the comprehensive feature vector using the machine learning model specifically comprises: The zero-order vibration frequency characteristic value and the differential adjustment frequency characteristic value are constructed into a comprehensive feature vector as the input of the machine learning model, to minimize the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability, as the training target of the model, based on the trained model, the drill string resonance risk probability value is output, wherein the machine learning model is a random forest model.

[0012] Preferably, the construction process of the machine learning model is: A plurality of sets of historical drilling zero sequence vibration frequency characteristic values, differential adjustment frequency characteristic values and drill string resonance risk probability values are obtained to construct a training set, a random forest extracts a plurality of subsample sets from the training set through bootstrap sampling, and an independent decision tree is constructed for each subsample set, each decision tree splits and judges based on the zero sequence vibration frequency characteristic values and the differential adjustment frequency characteristic values in the feature vector, and finally the probability prediction value of the drill string resonance risk under the current drilling state is obtained by weighted averaging the output results of all decision trees, and the structure and splitting rules of each tree in the forest are optimized in the training process to minimize the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability.

[0013] Preferably, the drill string is determined to be in a high-frequency resonance critical state according to the analysis result, specifically including: Whether the drill string resonance risk probability value is greater than or equal to a preset threshold value is judged, if yes, the drill string is in a high-frequency resonance critical state, and if not, the drill string is not in a high-frequency resonance critical state.

[0014] Preferably, the automatic triggering of the torque compensation mechanism dynamically adjusts the WOB, specifically including: When the drill string resonance risk probability value output by the machine learning model is greater than or equal to a preset threshold value, the system automatically starts the torque compensation control logic, adjusts the motor output torque curve and the WOB change rate in real time based on the change trend of the current zero sequence vibration frequency characteristic value and the differential adjustment frequency characteristic value, introduces the frequency offset strategy and the phase lag compensation mechanism, reduces the coupling strength between the excitation frequency and the natural frequency of the drill string, thereby inhibiting the concentrated release of torsional vibration energy at the weak part of the drill string, preventing the drill string fracture accident caused by high-frequency resonance, and realizing the closed-loop feedback control and intelligent early warning of the drilling process.

[0015] The beneficial effects of the present application are: (1) The present application constructs a multi-source heterogeneous signal fusion data perception system by collecting drilling string vibration frequency data and drilling pressure adjustment frequency data in real time during drilling, thereby realizing comprehensive monitoring of the drilling dynamic process. On this basis, the sparse Fourier transform (SFT) technology is innovatively introduced to analyze the vibration signal in the frequency domain, accurately extract the zero sequence vibration frequency component reflecting the overall rigid body vibration characteristics of the drilling string, and obtain the zero sequence vibration frequency characteristic value by calculating its time variation rate, which is used to quantify the vibration stability of the drilling string system. At the same time, aiming at the nonlinearity and time-varying characteristics of the drilling pressure adjustment behavior, the method of empirical mode decomposition (EMD) combined with Hilbert transform is used to carry out multi-level time-frequency analysis on the differential adjustment frequency sequence of drilling pressure, extract the instantaneous frequency information of each order intrinsic mode function, and construct the differential adjustment frequency characteristic value based on its standard deviation, which effectively represents the stability and fluctuation risk of the drilling pressure control process. Further, the above two types of characteristic values are constructed into a comprehensive feature vector, which is introduced into the intelligent recognition model trained based on the random forest algorithm as an input parameter, and through minimizing the error between the predicted resonance risk probability and the actual state, the high-precision discrimination of whether the drilling string is in the high-frequency resonance critical state is realized. Once the system determines that there is a resonance risk, an adaptive torque compensation mechanism is triggered, the motor output torque curve and the drilling pressure change rate are dynamically optimized, the frequency offset strategy and phase lag compensation mechanism are introduced, the coupling relationship between the excitation frequency and the natural frequency of the drilling string is actively broken, the concentrated release of torsional vibration energy in local structure is prevented, and thus the occurrence of malignant accidents such as drilling string fracture and downhole tool damage is effectively avoided, significantly improving the safety, stability and intelligent level of high-torque drilling operation under complex working conditions; (2) The present invention innovatively introduces a drill string resonance risk prediction model based on the random forest algorithm. By collecting multi-dimensional historical data, including zero-sequence vibration frequency eigenvalues, differential adjustment frequency eigenvalues, and corresponding drill string resonance occurrence state labels, a high-quality training sample set is constructed, and a self-service sampling method is used to generate multiple sub-sample sets, which are used to train decision tree-based models with diversity and differences. During the model training process, the objective function is to minimize the mean square error between the predicted drill string resonance risk probability value and the actual risk state, and the splitting rules and node judgment logic of each decision tree are optimized, thereby improving the generalization ability and anti-interference robustness of the overall model. After training, the random forest model can perform real-time evaluation of the resonance risk of the drill string system under the current drilling state, output statistically significant probabilistic prediction results, and provide a highly reliable judgment basis for the system control logic. Furthermore, this model and the automatically triggered torque compensation mechanism form a closed-loop response system: when the resonance risk probability exceeds a preset threshold, the control system quickly initiates a dynamic adjustment strategy. Combining the changing trends of the current zero-sequence vibration frequency and the differential adjustment frequency, it intelligently adjusts the motor output torque curve and the rate of change of the drilling pressure, implementing a frequency offset and phase lag compensation mechanism to effectively suppress the coupling effect between the excitation frequency and the drill string's natural frequency, breaking the system's positive feedback loop and preventing the continued accumulation of torsional vibration energy in weak areas of the drill string. The entire process achieves fully automated closed-loop control from data perception, feature extraction, intelligent identification, to adaptive control, significantly improving the system's response efficiency, control accuracy, and operational safety under high-torque drilling conditions. This provides solid technical support and an engineering practice foundation for the development of intelligent drilling technology under complex geological conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of the intelligent torque control method for high-torque drilling screws according to the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 As shown, the present invention is an intelligent torque control method applicable to a high-torque drilling screw, comprising the following steps: S1: Real-time acquisition of drill string vibration frequency data and WOB adjustment frequency data during drilling; S2: Extract the zero sequence vibration frequency component of the drill string from the vibration frequency data, calculate the zero sequence vibration frequency characteristic value according to the change rate of the zero sequence vibration frequency component, and use it to evaluate the vibration stability of the drill string; S3: Calculate the differential adjustment frequency based on the weight adjustment frequency data, calculate the differential adjustment frequency characteristic value according to the change degree of the differential adjustment frequency, and use it to evaluate the stability of the weight adjustment; S4: Construct the zero sequence vibration frequency characteristic value and the differential adjustment frequency characteristic value into a comprehensive feature vector, analyze the comprehensive feature vector using a machine learning model, and determine whether the drill string is in a high-frequency resonance critical state according to the analysis result; S5: If it is determined that the drill string is in a high-frequency resonance critical state, automatically trigger a torque compensation mechanism to dynamically adjust the weight on bit to ensure safe and stable operation of the drill string.

[0019] In S1, the vibration frequency data and the weight adjustment frequency data of the drill string are collected in real time during drilling, specifically including: During drilling, the vibration frequency data of the drill string is collected in real time, specifically including: installing high-precision three-axis vibration sensors on the drill string body, the vibration sensors are arranged at intervals along the axial direction of the drill string, for continuously monitoring the dynamic vibration response of the drill string during drilling; the vibration sensors collect acceleration time series signals of each part of the drill string, and convert the analog signals into digital signals through the built-in filtering and analog-to-digital conversion module, and the data acquisition system is used for unified time stamp marking and storage, forming a complete drill string vibration frequency data sequence.

[0020] Further, the acquisition of the weight adjustment frequency data includes: integrating pressure sensors and speed encoders in the output end of the drilling motor and the top drive system, and collecting the instantaneous weight on bit value and its change frequency applied to the drill string during the weight adjustment process in real time; the data acquisition system performs sliding window differential processing on the weight signal, extracts the weight fluctuation period within a unit time, calculates the time series data of the weight adjustment frequency, and synchronously archives it with the vibration frequency data for subsequent feature extraction and resonance risk assessment.

[0021] In S2, the zero sequence vibration frequency component of the drill string is extracted from the vibration frequency data, and the zero sequence vibration frequency characteristic value is calculated according to the change rate of the zero sequence vibration frequency component, which is used to evaluate the vibration stability of the drill string, specifically including: During drilling, the vibration frequency data of the drill string is obtained, the zero sequence vibration frequency component of the drill string is extracted from the vibration frequency data, and the zero sequence vibration frequency characteristic value is calculated according to the change rate of the zero sequence vibration frequency component, specifically including: The time series of the obtained vibration frequency data of the drill string is applied to sparse Fourier transform to obtain its frequency domain sparse representation: ; In the formula, denotes a sparse spectrum result, x n denotes a time series of vibration frequency data, n denotes a number of collection points, k denotes a number of main frequency components for a set sparsity, denotes a frequency component; identifies a zero sequence vibration frequency component from the sparse spectrum and calculates a rate of change of the corresponding frequency component over time: ; wherein, denotes a zero sequence vibration frequency component, t j denotes a collection time point, j denotes a number of collection time points, j denotes a number of collection time points, denotes a zero sequence vibration frequency value extracted at the collection time point, j denotes a rate of change of the zero sequence vibration frequency value at the collection time point; According to the rate of change of the zero sequence vibration frequency value, a zero sequence vibration frequency characteristic value is calculated, and the calculation expression is: j ; wherein, zv denotes a zero sequence vibration frequency characteristic value, denotes a total number of collection time points. F m In the drilling process, it is judged whether the zero sequence vibration frequency characteristic value is greater than or equal to a preset threshold value. If yes, the vibration of the drill string is unstable, and if no, the vibration of the drill string is stable.

[0022] In the drilling process, it is judged whether the zero sequence vibration frequency characteristic value is greater than or equal to a preset threshold value. If yes, the vibration of the drill string is unstable, and if no, the vibration of the drill string is stable.

[0023] ​​It should be noted that: the application extracts the main frequency component in the frequency domain by collecting the drilling string vibration frequency data time series, and identifies the zero sequence vibration frequency representing the overall rigid body vibration mode of the drilling string from the main frequency component; further, the rate of change of the frequency with time is calculated, and the zero sequence vibration frequency characteristic value is obtained by accumulating the absolute value of the rate of change, which is used to quantify the stability of the drilling string vibration state; when the characteristic value exceeds the preset threshold value, it is determined that the drilling string vibration tends to be unstable, and there may be a resonance risk. The advantages of the technical scheme of the application are: the application has higher calculation efficiency and frequency resolution, and is especially suitable for feature extraction of non-stationary and short-time signals; combined with the dynamic evaluation mechanism of the zero sequence frequency change rate, the vibration abnormal trend of the drilling string system can be identified earlier and more accurately, and reliable basis is provided for subsequent intelligent early warning and control. The innovation lies in that the sparse Fourier transform is first introduced into the drilling vibration monitoring field, and a vibration stability evaluation index system with the zero sequence frequency characteristic value as the core is constructed, which improves the perception ability and response accuracy of the resonance risk in the high-torque drilling process.

[0024] In S3, the differential adjustment frequency is calculated based on the drilling pressure adjustment frequency data, and the differential adjustment frequency characteristic value is calculated according to the change degree of the differential adjustment frequency, which is used to evaluate the stability of the drilling pressure adjustment, specifically including: In the drilling process, the drilling pressure adjustment frequency data is obtained, the differential adjustment frequency is calculated based on the drilling pressure adjustment frequency data, and the differential adjustment frequency characteristic value is calculated according to the change degree of the differential adjustment frequency, specifically including: The differential adjustment frequency of the time series of the drilling pressure adjustment frequency data is calculated, specifically the difference between adjacent time points is calculated to obtain the differential adjustment frequency, all differential adjustment frequencies are integrated into a differential adjustment frequency time series, the differential adjustment frequency time series is subjected to empirical mode decomposition to obtain a group of intrinsic mode functions, and further Hilbert transform is performed on each mode function to obtain the instantaneous differential adjustment frequency, and the calculation expression is: In the formula, t represents the time series collection point, ω i (t) represents the instantaneous differential adjustment frequency of the i th intrinsic mode function obtained after Hilbert transform, represents Hilbert transform, a represents an imaginary unit, i represents the number of intrinsic mode functions, and arg represents the amplitude angle of a complex number, represents derivative operation, and IMF represents intrinsic mode function; The standard deviation of the instantaneous differential adjustment frequency of all intrinsic mode functions is calculated to obtain the differential adjustment frequency characteristic value.

[0025] In the drilling process, it is judged whether the differential adjustment frequency characteristic value is greater than or equal to the preset threshold value, if yes, the drilling pressure adjustment is unstable, and if no, the drilling pressure adjustment is stable. ​

[0026] It should be noted that: the present application carries out time series difference processing on the real-time collected drilling pressure adjustment frequency data, to obtain a difference adjustment frequency sequence; then the empirical mode decomposition is used to decompose it into a plurality of intrinsic mode functions, and then the Hilbert transform is performed on each order intrinsic mode function to obtain its instantaneous frequency information, and the standard deviation of each order frequency is calculated, and finally the difference adjustment frequency characteristic value is constructed by weighted average, which is used to quantify the dynamic stability in the drilling pressure adjustment process. The technical scheme can effectively capture the small but key frequency fluctuations in the drilling pressure adjustment process by combining nonlinear and non-stationary signal processing methods, improve the sensitivity and recognition accuracy of the system to changes in the excitation source; the present application constructs a quantitative criterion with the difference adjustment frequency characteristic value as the core, realizes the fine monitoring of the drilling pressure control behavior, thereby providing more accurate risk judgment basis for intelligent torque control, and significantly improves the safety and reliability of high torque drilling operation.

[0027] In S4, the zero sequence vibration frequency characteristic value and the difference adjustment frequency characteristic value are constructed into a comprehensive feature vector, and a machine learning model is used to analyze the comprehensive feature vector, and according to the analysis result, it is determined whether the drill string is in a high-frequency resonance critical state, which specifically includes: The zero sequence vibration frequency characteristic value and the difference adjustment frequency characteristic value are constructed into a comprehensive feature vector as the input of the machine learning model, and the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability is minimized as the training target of the model, and based on the trained model, the drill string resonance risk probability value is output, wherein the machine learning model is a random forest model.

[0028] A plurality of sets of historical drilling zero sequence vibration frequency characteristic values, difference adjustment frequency characteristic values and drill string resonance risk probability values are obtained, and are constructed into a training set, and the random forest extracts a plurality of sub-sample sets from the training set through bootstrap sampling, and constructs an independent decision tree for each sub-sample set, each decision tree splits and judges based on the zero sequence vibration frequency characteristic value and the difference adjustment frequency characteristic value in the feature vector, and finally the weighted average of all decision tree output results is obtained. The probability prediction value of the drill string resonance risk under the current drilling state, the training process minimizes the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability as the objective function, and optimizes the structure and splitting rule of each tree in the forest.

[0029] After the training is completed, the real-time zero sequence vibration frequency characteristic value and the differential adjustment frequency characteristic value are constructed into a comprehensive feature vector, which is input into the trained random forest model, and the drill string resonance risk probability value of the drill string system under the current drilling state is output. The drill string resonance risk probability value can be used as an important basis for subsequent control strategy adjustment. When the risk probability exceeds the preset threshold, the system automatically triggers the dynamic adjustment mechanism of torque and weight on bit, realizing closed-loop early warning and active inhibition of drill string resonance risk.

[0030] In S5, if it is determined that the drill string is in a high-frequency resonance critical state, a torque compensation mechanism is automatically triggered to dynamically adjust the weight on bit, ensuring the safe and stable operation of the drill string, specifically including: During drilling, when the system determines that the drill string is in a high-frequency resonance critical state, that is, the drill string resonance risk probability value output by the machine learning model is greater than or equal to the preset threshold, the torque compensation mechanism is automatically triggered. The torque compensation mechanism sends dynamic adjustment instructions to the drilling motor and the weight on bit adjustment device through the control module, specifically including the following control strategies: First, based on the change trend of the current zero sequence vibration frequency characteristic value and the differential adjustment frequency characteristic value, the system real-time evaluates the vibration energy accumulation rate and excitation frequency offset direction of the drill string system; second, according to the evaluation result, the motor output torque curve is dynamically adjusted, and a nonlinear gradual torque reduction strategy is adopted to avoid additional impact on the drill string caused by instantaneous large torque fluctuations; at the same time, the control system synchronously adjusts the weight on bit change rate, introduces a frequency offset strategy to make the excitation frequency actively deviate from the natural vibration frequency of the drill string, and reduces the resonance coupling possibility; in addition, a phase lag compensation mechanism is added to the control logic, which applies a small time delay to the weight on bit and torque adjustment actions to break the positive feedback cycle of the system and further weaken the continuous accumulation of torsional vibration energy; the whole process continuously monitors the drill string vibration state and control effect through a closed-loop feedback mechanism, and dynamically optimizes the control parameters according to the real-time acquisition data to ensure that the drill string can still maintain safe and stable operation under the high-frequency resonance critical state, preventing the occurrence of downhole accidents such as drill string rupture and tool damage caused by resonance.

[0031] The present application also includes an intelligent torque control method experimental data comparison table suitable for high-torque drilling screws:

[0032] Data analysis and optimization effect summary: Torque stability: The intelligent torque control method reduces the average torque fluctuation to 50 Nm and the maximum torque peak to 200 Nm, significantly improving the stability and controllability of the torque.

[0033] Drilling efficiency: Through more accurate torque control, the drilling speed is increased from 15 meters per hour to 20 meters per hour, an increase of 33.3%, and the drilling period is shortened by 25%.

[0034] Device durability: tool wear rate reduced from 10% to 4%, significantly extending the service life of drill bits and other key components, reducing maintenance costs.

[0035] Energy saving: intelligent control system reduces unit energy consumption by 25%, helping to save energy and reduce emissions, and improving energy utilization efficiency.

[0036] System response capability: system response time is shortened from 30 seconds to 10 seconds, significantly enhancing the rapid response capability to underground environment changes, ensuring the safety and efficiency of the drilling process.

[0037] Drilling success rate: drilling success rate increased to 95%, and accident risk decreased from 0.05 times per drilling to 0.01 times, indicating that the intelligent torque control method has significant advantages in improving operation safety.

[0038] The intelligent torque control method for high-torque drilling screw provided by the embodiment of the present application realizes intelligent identification of whether the drill string is in a high-frequency resonance critical state by real-time acquisition of drill string vibration frequency data and drilling pressure adjustment frequency data, extraction of zero sequence vibration frequency eigenvalue combined with sparse Fourier transform, and analysis of differential adjustment frequency by empirical mode decomposition and Hilbert transform. When it is determined that there is a resonance risk, the system automatically triggers a torque compensation mechanism to dynamically adjust the drilling pressure and torque output, ensuring the safe and stable operation of the drill string.

[0039] Working principle of the present application: The present application aims to solve the technical problems of high-frequency resonance caused by the mismatch between drill string vibration and weight adjustment, leading to drill tool damage and even downhole accidents. The technical solution of the present application includes the following core steps: First, in the drilling process, through the multi-point three-axis vibration sensor installed on the drill string body and the pressure sensor and rotational speed encoder at the drill motor end, real-time acquisition of drill string vibration frequency data and weight adjustment frequency data is carried out; second, sparse Fourier transform (SFT) is used to extract the zero sequence vibration frequency component from the vibration frequency data, and the characteristic value of the zero sequence vibration frequency corresponding to the change rate is calculated, which is used to quantify the overall vibration stability of the drill string; at the same time, the difference processing is carried out on the weight adjustment frequency time series, and the instantaneous frequency information is extracted by combining empirical mode decomposition (EMD) and Hilbert transform, and the difference adjustment frequency characteristic value is further calculated to evaluate the stability of the weight adjustment behavior; then, the above two characteristic values are constructed into a comprehensive feature vector, which is input into the machine learning model trained based on the random forest algorithm to predict whether the drill string system is in a high-frequency resonance critical state; if it is determined that there is a resonance risk, the torque compensation mechanism is automatically triggered, the motor output torque and the weight change rate are dynamically adjusted, the frequency offset and phase lag compensation strategy is introduced, the coupling relationship between the excitation frequency and the natural frequency is broken, and the concentrated release of torsional vibration energy is suppressed. The present application combines signal processing, feature extraction and machine learning modeling technology to build a closed-loop intelligent torque control system, realizes early identification and active intervention of drill string resonance risk, and significantly improves the safety, stability and intelligent level of high-torque drilling operation.

[0040] The above formulas are dimensionless numerical calculations, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the latest real situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0041] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0042] It should be understood that the term "and / or" herein merely describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents that the associated objects before and after are an "or" relationship, but can also represent an "and / or" relationship, which can be understood according to the context before and after.

[0043] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of the processes should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0044] The above describes one embodiment of the present application in detail, but the content described is only the preferred embodiment of the present application, and cannot be considered as limiting the scope of the implementation of the present application. Any equivalent changes and improvements made within the scope of the present application are still within the scope of the patent coverage of the present application.

Claims

1. An intelligent torque control method for high-torque drilling screws, characterized in that: The following steps are involved: S1: Real-time acquisition of drill string vibration frequency data and WOB adjustment frequency data during drilling; S2: Extract the zero-sequence vibration frequency component of the drill string from the vibration frequency data. Calculate the zero-sequence vibration frequency characteristic value based on the change rate of the zero-sequence vibration frequency component to evaluate the vibration stability of the drill string. S3: Calculate the differential adjustment frequency based on the WOB adjustment frequency data, and calculate the differential adjustment frequency characteristic value according to the degree of change of the differential adjustment frequency to evaluate the stability of the WOB adjustment; S4: constructing a comprehensive feature vector from the zero-sequence vibration frequency eigenvalue and the differential adjustment frequency eigenvalue, analyzing the comprehensive feature vector using a machine learning model, and determining whether the drill string is in a high-frequency resonance critical state based on the analysis results; S5: If the drill string is determined to be in a critical state of high-frequency resonance, the torque compensation mechanism is automatically triggered to dynamically adjust the drilling pressure to ensure safe and stable operation of the drill string.

2. The intelligent torque control method for high-torque drilling screw according to claim 1, characterized in that: The process of obtaining the zero-sequence vibration frequency characteristic value is as follows: During the drilling process, the vibration frequency data of the drill string is obtained, the zero-sequence vibration frequency component of the drill string is extracted from the vibration frequency data, and the zero-sequence vibration frequency characteristic value is calculated according to the change rate of the zero-sequence vibration frequency component. Specifically, the following steps are performed: Apply sparse Fourier transform to the time series of the drill string vibration frequency data to obtain its sparse representation in the frequency domain: ; Where, represents the sparse spectrum result, x ( n ) represents the time series of vibration frequency data, n Indicates the number of collection points, k is the set sparsity, indicating the number of main frequency components, represents the frequency component; Identifying zero-sequence vibration frequency components from sparse spectrum , and calculate the rate of change of the corresponding frequency component over time: ; Where, Represents the zero-sequence vibration frequency component, t j Indicates the j Collection time points, j Indicates the number of acquisition time points, Indicates the j The zero-sequence vibration frequency value extracted at each acquisition time point, Indicates the j The rate of change of the zero-sequence vibration frequency value at each acquisition time point; According to the rate of change of the zero-sequence vibration frequency value, the zero-sequence vibration frequency characteristic value is calculated, and the calculation expression is: ; Where, F zv represents the zero-sequence vibration frequency characteristic value, m Indicates the total number of acquisition time points.

3. The intelligent torque control method for high-torque drilling screw according to claim 2, characterized in that: The evaluation of the vibration stability of the drill string specifically includes: During the drilling process, it is determined whether the zero-sequence vibration frequency characteristic value is greater than or equal to a preset threshold. If so, the vibration of the drill string is unstable; otherwise, the vibration of the drill string is stable.

4. The intelligent torque control method for high-torque drilling screw according to claim 1, characterized in that: The process of obtaining the differential adjustment frequency characteristic value is as follows: During the drilling process, the drilling pressure adjustment frequency data is obtained, the differential adjustment frequency is calculated based on the drilling pressure adjustment frequency data, and the differential adjustment frequency characteristic value is calculated according to the degree of change of the differential adjustment frequency, specifically including: The differential adjustment frequency of the time series of the WOB adjustment frequency data is calculated. Specifically, the difference between adjacent time points is calculated to obtain the differential adjustment frequency. All differential adjustment frequencies are integrated into a differential adjustment frequency time series. The differential adjustment frequency time series is subjected to empirical mode decomposition to obtain a set of intrinsic mode functions. The Hilbert transform is further performed on each mode function to obtain the instantaneous differential adjustment frequency. The calculation expression is: ; Where t represents the time series acquisition point, ω i (t) represents the instantaneous differential adjustment frequency of the i-th intrinsic mode function obtained after Hilbert transform, represents the Hilbert transform, a represents the imaginary unit, i represents the number of intrinsic mode functions, arg represents the argument of the complex number, represents the derivative operation, IMF represents the intrinsic mode function; The standard deviation of the instantaneous differential adjustment frequencies of all natural mode functions is calculated to obtain the differential adjustment frequency eigenvalues.

5. The intelligent torque control method for high-torque drilling screw according to claim 4, characterized in that: The evaluation of the stability of the bit weight adjustment specifically includes: During the drilling process, it is determined whether the differential adjustment frequency characteristic value is greater than or equal to a preset threshold. If so, the drilling pressure adjustment is unstable; otherwise, the drilling pressure adjustment is stable.

6. The intelligent torque control method for high-torque drilling screw according to claim 1, characterized in that: The use of a machine learning model to analyze the comprehensive feature vector specifically includes: The zero-sequence vibration frequency eigenvalue and the differential adjustment frequency eigenvalue are constructed into a comprehensive eigenvector as the input of a machine learning model to minimize the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability. Based on the trained model, the drill string resonance risk probability value is output, wherein the machine learning model is a random forest model.

7. The intelligent torque control method for high-torque drilling screw according to claim 6, characterized in that: The process of building the machine learning model is as follows: Multiple sets of historical drilling zero-sequence vibration frequency eigenvalues, differential adjustment frequency eigenvalues, and drill string resonance risk probability values ​​are obtained to construct a training set. Random forest extracts multiple subsample sets from the training set through self-service sampling, and constructs an independent decision tree for each subsample set. Each decision tree makes a split judgment based on the zero-sequence vibration frequency eigenvalue and differential adjustment frequency eigenvalue in the eigenvector. Finally, the probability prediction value of the drill string resonance risk under the current drilling state is obtained by taking a weighted average of the output results of all decision trees. During the training process, the objective function is to minimize the error between the predicted drill string resonance risk probability value and the actual drill string resonance risk probability, and to optimize the structure and splitting rules of each tree in the forest.

8. The intelligent torque control method for high-torque drilling screw according to claim 7, characterized in that: Determining whether the drill string is in a high-frequency resonance critical state based on the analysis results specifically includes: It is determined whether the drill string resonance risk probability value is greater than or equal to a preset threshold value. If so, the drill string is in a high-frequency resonance critical state; otherwise, the drill string is not in a high-frequency resonance critical state.

9. The intelligent torque control method for high-torque drilling screw according to claim 1, characterized in that: The automatic triggering of the torque compensation mechanism to dynamically adjust the drilling pressure specifically includes: When the drill string resonance risk probability value output by the machine learning model is greater than or equal to the preset threshold, the system automatically starts the torque compensation control logic. Based on the changing trends of the current zero-sequence vibration frequency characteristic value and the differential adjustment frequency characteristic value, it adjusts the motor output torque curve and the drilling pressure change rate in real time, introduces a frequency offset strategy and a phase lag compensation mechanism, reduces the coupling strength between the excitation frequency and the drill string natural frequency, thereby suppressing the concentrated release of torsional vibration energy in weak parts of the drill string, preventing drill string fracture accidents caused by high-frequency resonance, and realizing closed-loop feedback control and intelligent early warning of the drilling process.

Images