Petroleum drilling parameter optimization method and system and storage medium

Abstract

The invention relates to the technical field of drilling parameter optimization control, and provides a petroleum drilling parameter optimization method and system and a storage medium. Comprising the steps that torque fluctuation amplitude and bit pressure response delay data are collected in real time, an original data sequence is constructed, and parameter trend data are obtained; analyzing the parameter trend data to identify a stratum change mode, extracting trend feature information, and when the trend feature information meets a preset condition, determining a corresponding abnormal mode identifier; generating a parameter adjustment direction of drilling control according to the abnormal mode identifier, and optimizing the parameter adjustment direction in combination with parameter weight analysis and a real-time feedback mechanism to obtain final optimization guidance data for guiding drilling operation; and mapping the final optimization guidance data to a control interface, and generating a dynamic adjustment instruction for drilling operation. The problems that in the existing petroleum drilling process, parameter adjustment lags behind, and stratum changes are difficult to adapt are solved, and intelligent optimization and dynamic control over drilling parameters are achieved.

Description

Technical Field

[0001] This invention relates to the field of drilling parameter optimization technology, and in particular to a method, system and storage medium for optimizing oil drilling parameters. Background Technology

[0002] As a pillar technology in the energy development field, the success of oil drilling engineering is not only related to the economics of resource acquisition, but also directly to the stability of the operating platform and environmental protection requirements. In actual operations, drilling equipment must cope with the variable characteristics of underground formations, such as nonlinear jumps in hardness index, sudden anomalies in pressure structure, and regional geological differences. These factors often lead to a sharp increase in torque fluctuation or a prolonged delay in drilling pressure response, which in turn can cause hidden dangers such as drill bit jamming, trajectory deviation, or equipment overload. Traditional methods mostly rely on empirical parameter settings or static control models, such as simple feedback loops based on preset thresholds. Although these methods can maintain basic stability in homogeneous formations, they exhibit significant response lag issues when facing the deep and complex environment of oil wells. The raw dynamic data collected by sensors is easily interfered with by high-frequency noise, making it impossible to accurately extract formation change trends. This causes parameter adjustments to deviate from the real-time formation model, resulting in a decrease in drilling efficiency of about 15%-20% and amplifying the risk of downhole accidents.

[0003] Current technologies are often limited to single-parameter monitoring, such as tracking only torque fluctuations while ignoring the coupling effect of drilling pressure delay, or using fixed filtering algorithms to process signals. These methods struggle to adapt to the rapid transition of formation hardness from soft soil to hard rock in oil drilling. Such methods easily filter out low-frequency trend features during signal processing, leading to delays in anomaly pattern recognition. Consequently, the generated parameter adjustment directions lack coordinated optimization; for example, there is a priority conflict between rotation speed correction and motion trajectory angle, making adaptive fusion through weight analysis impossible. Ultimately, this amplifies equipment wear and prolongs drilling cycles. This disconnect between perception and response is particularly pronounced in deep well or offshore platform scenarios. Historical data shows that parameter misalignment accidents caused by sudden formation changes account for as much as 30%, highlighting the urgent need for a method that integrates real-time data acquisition, trend feature extraction, and closed-loop optimization to bridge this gap.

[0004] Therefore, this invention is proposed. By constructing original sequences with time stamps, analyzing parameter trend data after filtering out interference, and generating anomaly markers and optimizing weights based on a historical pattern library, it achieves accurate capture of formation changes and dynamic parameter adjustment. This not only solves the pain points of response lag and inaccurate control in existing technologies, but also significantly improves the robustness and economy of drilling operations. Summary of the Invention

[0005] This invention provides a method, system, and storage medium for optimizing oil drilling parameters, which is used for intelligent optimization and dynamic control mechanism of drilling parameters for complex formation conditions. Overall, it achieves a closed-loop technical effect from anomaly identification, decision generation to execution feedback.

[0006] In a first aspect, the present invention provides a method for optimizing oil drilling parameters, the method comprising: Step S1: Collect torque fluctuation amplitude and drilling pressure response delay data during the oil drilling process using sensors, and construct a raw sequence with time stamps; perform filtering and smoothing processing on the raw sequence to obtain parameter trend data reflecting formation characteristics; Step S2: Based on the parameter trend data, analyze the stratigraphic change pattern and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, then determine the abnormal pattern identifier by comparing with the historical bottom database. Step S3: For the abnormal mode identifier, generate the parameter adjustment direction for drilling control; optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least rotation speed and motion trajectory deflection angle; Step S4: Based on the final optimization guidance data, map it to the control interface and generate dynamic adjustment instructions.

[0007] As a preferred embodiment of the present invention, step S1, constructing an original sequence with a time identifier, includes: Torque fluctuation amplitude and drill pressure response delay data during oil drilling are acquired in real time using a sensor array. These data are then arranged chronologically to construct an initial dynamic parameter sequence. The initial dynamic parameter sequence is preliminarily labeled based on the formation hardness index, generating a raw acquisition sequence with timestamps. This raw acquisition sequence is stored as input data for subsequent processing, ensuring that each data point in the raw acquisition sequence corresponds to a unique time identifier. The timestamps of the raw acquisition sequence are verified; if any timestamps are missing, they are supplemented using interpolation methods to obtain a complete raw sequence with timestamps.

[0008] As a preferred embodiment of the present invention, step S1, obtaining parameter trend data reflecting formation characteristics, includes: The original sequence is denoised by signal smoothing; high-frequency interference from torque fluctuation amplitude in the original sequence is filtered out by a filtering algorithm; low-frequency variation characteristics related to formation hardness index are retained; and a smoothed parameter trend curve is generated based on the low-frequency variation characteristics. The smoothed parameter trend curve is verified to ensure that it reflects the changing pattern of formation characteristics. If the parameter trend curve has abnormal fluctuations, it is further optimized by secondary smoothing to obtain the final parameter trend data that reflects the formation characteristics.

[0009] As a preferred embodiment of the present invention, step S2, extracting the corresponding trend feature information, includes: Based on the parameter trend data, a continuous time window is selected for analysis; the trend vector mapping within the continuous time window is calculated; combined with the stratigraphic type identifier, the frequency of potential stratigraphic change patterns in the trend vector mapping is analyzed; based on the frequency of stratigraphic change patterns, trend feature vectors reflecting stratigraphic influence are extracted; the trend feature vectors are standardized, and by analyzing the rate of change of the trend feature vectors, key time points of stratigraphic change are determined, and corresponding trend feature information is generated.

[0010] As a preferred embodiment of the present invention, in step S2, if the trend feature information meets preset conditions, an abnormal pattern identifier is determined by comparing historical data, including: Calculate the absolute value of the slope of the trend feature information; if the absolute value of the slope exceeds the upper limit of the preset adjustment value, it is determined that the trend feature information meets the preset conditions; extract the expected adjustment direction matching the trend feature information from the pre-established stratigraphic model library by using historical data comparison method; determine the corresponding abnormal change pattern identifier according to the expected adjustment direction; and verify the matching degree of the abnormal change pattern identifier.

[0011] As a preferred embodiment of the present invention, in step S3, a preliminary parameter adjustment direction is generated for the abnormal mode identifier, including: Based on the abnormal pattern identifier, a geological change description with high pattern matching accuracy is obtained; the geological change description is analyzed by combining the optimization target priority of rotation speed correction and motion trajectory deflection angle; based on the analysis results, a preliminary parameter adjustment direction sequence is generated; the feasibility of the parameter adjustment direction sequence is verified, and if the parameter adjustment direction sequence exceeds the working range, the parameter adjustment direction sequence is corrected through constraint conditions to obtain an adjustment direction that meets the requirements.

[0012] As a preferred embodiment of the present invention, step S3, obtaining the final optimization guidance data, includes: The parameter adjustment direction sequence is analyzed by parameter linkage weight analysis; the adjustment value time window is dynamically divided by combining the real-time feedback cycle; the parameter adjustment weight in each time window is calculated according to the dynamically divided time window; the parameter adjustment direction sequence is optimized and adjusted according to the parameter adjustment weight; the final drilling parameter optimization guidance sequence is generated, and the stability of the final drilling parameter optimization guidance sequence is verified.

[0013] As a preferred embodiment of the present invention, step S4, generating a dynamic adjustment instruction, includes: The final optimized guidance data is input into the navigation control interface; parameter mapping is performed for rotation speed correction and motion trajectory deviation angle; a real-time parameter fine-tuning scheme is generated based on the parameter mapping results; dynamic adjustment instructions for drilling operations are determined through the real-time parameter fine-tuning scheme; the dynamic adjustment instructions are prioritized, with adjustments to key parameters being executed first; the execution effect of the dynamic adjustment instructions is monitored in real time to confirm whether the dynamic adjustment instructions meet the expected goals.

[0014] Secondly, the present invention also provides an oil drilling parameter optimization system for implementing the above-mentioned method, the system comprising: The data acquisition unit is used to collect data on torque fluctuation amplitude and drilling pressure response delay during the oil drilling process through sensors, and to construct a raw sequence with time stamps. The data preprocessing unit is used to filter and smooth the original sequence to obtain parameter trend data reflecting the formation characteristics; An anomaly identification unit is used to analyze the stratigraphic change pattern based on the parameter trend data and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, the anomaly pattern identifier is determined by comparison with the historical bottom database. The parameter adjustment and optimization unit is used to generate the parameter adjustment direction for drilling control based on the abnormal mode identifier; and to optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least rotation speed and motion trajectory deflection angle. The instruction generation unit is used to map the final optimization guidance data to the control interface and generate dynamic adjustment instructions.

[0015] Thirdly, the present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the above-described method.

[0016] The beneficial effects of this invention are as follows: This invention, through continuous acquisition and time-series processing of multi-source dynamic parameters during drilling, combined with signal denoising, trend extraction, and feature analysis, can stably and accurately depict the true trend of formation characteristics changing over time, fundamentally improving data quality and the ability to perceive formation changes. By comparing and analyzing with a historical formation pattern database, it can promptly identify abnormal change patterns matching the current working conditions, enabling the system to respond at the initial stage of formation abrupt changes and avoiding the risks associated with traditional reliance on human experience or delayed judgment. For identified abnormal patterns, by introducing parameter optimization target priorities and parameter linkage logic, it conducts targeted analysis of formation changes and generates structured data. The parameter adjustment direction sequence provides a clear basis and inherent order for adjusting key parameters such as rotation speed and trajectory deflection, thus avoiding control instability caused by disordered adjustments of multiple parameters. Simultaneously, by verifying the feasibility and correcting constraints of parameter adjustment directions, the system ensures that all adjustment schemes remain within the safe operating range of the drilling equipment and working environment, improving the engineering feasibility of the schemes. After the parameter decision-making layer is completed, the system maps the optimized guidance data to the navigation control interface. Through parameter mapping, fine-tuning scheme generation, and command priority sorting, the abstract optimization results are transformed into directly executable dynamic adjustment commands, enabling the drilling equipment to achieve continuous and smooth parameter adjustments according to the rhythm of formation changes. At the same time, by monitoring the execution effect in real time and introducing a feedback mechanism, the adjustment results are used to influence subsequent decision-making processes, forming an adaptive optimization closed loop. Through the synergy of the above technical solutions, accurate identification of formation changes, intelligent decision-making and safe execution of drilling parameters are achieved, effectively improving the real-time performance, stability, and safety of parameter adjustments during oil drilling, reducing equipment wear and operational risks, and significantly enhancing the overall efficiency and reliability of drilling operations under complex formation conditions, demonstrating significant engineering application value. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of an oil drilling parameter optimization method in an embodiment; Figure 2 This is a schematic diagram of the dynamic time window division method in the embodiment; Figure 3 This is a schematic diagram illustrating the stability verification effect in the embodiment; Figure 4 This is an example of an oil drilling parameter optimization system structure. Detailed Implementation

[0019] This invention provides a method, system, and storage medium for optimizing oil drilling parameters. The terms "first," "second," "third," "fourth," etc. (if applicable) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] In complex underground formation environments, traditional drilling control often relies on static parameter settings or empirical rules, which cannot respond in real time to dynamic factors such as nonlinear jumps in formation hardness index and abrupt changes in pressure structure. This leads to a sharp increase in torque fluctuations or prolonged drilling pressure response delays, causing potential hazards such as drill bit jamming, trajectory deviation, and equipment overload, thereby amplifying downhole accident risks and reducing drilling efficiency. Although existing methods can collect raw dynamic data, the signal processing stage is susceptible to high-frequency noise interference, making it difficult to extract reliable parameter trend data and formation change patterns. This results in delayed anomaly identification and a lack of coordinated optimization of parameter adjustments (such as drilling speed and trajectory deviation angle), ultimately causing response disconnect and economic losses. This invention achieves accurate perception and adaptive adjustment of formation characteristics through a closed-loop process of constructing raw sequences with time stamps, filtering and smoothing trend analysis, generating anomaly pattern identifiers by comparing with historical databases, and weight optimization and interface mapping. This significantly improves operational robustness and safety, providing key support for the intelligent transformation of oil resource development.

[0021] For ease of understanding, the specific process of the embodiments of the present invention will be described below, such as... Figure 1 As shown in the figure, an oil drilling parameter optimization method in an embodiment of the present invention includes: Step S1: Collect torque fluctuation amplitude and drilling pressure response delay data during the oil drilling process using sensors, and construct a raw sequence with time stamps; perform filtering and smoothing processing on the raw sequence to obtain parameter trend data reflecting formation characteristics; In step S1, constructing the original sequence with time identifiers includes: The torque fluctuation amplitude and drilling pressure response delay data during the oil drilling process are acquired in real time using a sensor array. The torque fluctuation amplitude and drilling pressure response delay data are arranged chronologically to construct an initial dynamic parameter sequence. The initial dynamic parameter sequence is preliminarily labeled according to the formation hardness index to generate a timestamped original acquisition sequence. The timestamp of the original acquisition sequence is verified; if a timestamp is missing, it is completed using interpolation to obtain a timestamped original sequence.

[0022] Specifically, in one embodiment, torque fluctuation amplitude and drilling pressure response delay data during the oil drilling process are acquired in real time using a sensor array, and a time-stamped raw sequence is constructed. Specifically, a multi-channel sensor array deployed on the drilling equipment, such as torque sensors and pressure response sensors, continuously monitors the real-time dynamics of the drilling. Torque fluctuation amplitude data reflects the mechanical response characteristics of the drill bit's interaction with the formation, while drilling pressure response delay data captures the time lag effect of drilling fluid pressure transmission to the drill bit, quantifying the dynamic characteristics of formation permeability and pressure transmission. Increased delay often indicates abrupt changes in formation structure. The two types of data are temporally correlated during acquisition to ensure spatiotemporal consistency in the subsequent sequence construction. Subsequently, the torque fluctuation amplitude and the drill pressure response delay data are arranged according to the time sequence of acquisition to form an initial dynamic parameter sequence. At this time, each data point in the sequence is initially indexed by timestamp, thus laying the foundation for the time-series analysis of formation changes. Among them, the torque fluctuation amplitude data is the change amplitude of torque value of the drill bit affected by formation resistance during drilling. It directly reflects the real-time feedback of the formation hardness index. For example, when the drill bit encounters a hard rock layer, the torque fluctuation amplitude will increase sharply (a significant increase from the baseline value). This is due to the nonlinear jump of formation friction and shear strength, which leads to increased drilling resistance. If there is no timely response, it may cause drill bit overload or jamming risk.

[0023] The initial dynamic parameter sequence was preliminarily labeled based on the formation hardness index to generate a timestamped raw acquisition sequence. During the labeling process, the formation hardness index was used as a quantitative indicator of formation physical properties. It showed a positive correlation with torque fluctuation amplitude (i.e., torque fluctuation amplitude tends to amplify as hardness increases) and a negative correlation with drill pressure response delay (i.e., delay usually shortens as hardness increases). Specifically, in hard formations, the low porosity and high elastic modulus of the rock make the pressure wave propagation path more "rigid," reducing the buffering and diffusion effects of the medium, leading to a shorter delay time. The labeling process involved matching each data point in the sequence with the corresponding formation hardness index within a preset formation hardness index threshold range, such as a grading interval for different formation types. For example, data points with larger torque fluctuation amplitudes were associated with higher hardness intervals, thus achieving data point classification mapping. Based on this, labels were added to the matched data points. Labels such as "soft formation" or "hard formation" are used, and precise timestamps are added to each labeled point. This ensures a multi-dimensional correspondence between torque fluctuation amplitude data, drill pressure response delay data, and formation hardness index labels in the sequence. That is, each timestamp contains a multi-field structure including torque value, delay value, and hardness index label. Finally, a complete raw acquisition sequence with timestamps is generated. This labeling mechanism not only strengthens the logical connection between data, such as the causal relationship between torque fluctuation and hardness index, but also provides semantic guidance for subsequent signal processing. For example, in the scenario of alternating soft and hard formations, the labeling can highlight the gradual increase trend of torque fluctuation amplitude in hard formations, thereby improving the accuracy of formation change identification in the overall method. In the scenario of deep well drilling, the formation hardness index can be calibrated through historical drilling data to further optimize the adaptability of the labeling, ensuring that the sequence captures subtle changes at formation transition points and avoids potential risks such as drill bit jamming.

[0024] Next, the original acquisition sequence is stored as input data for subsequent processing. This ensures that each data point in the original acquisition sequence corresponds to a unique time identifier, which serves as the primary key. This ensures the spatiotemporal alignment of torque fluctuation amplitude data and drilling pressure response delay data, avoiding parameter deviations caused by time misalignment during the optimization process. Simultaneously, the time identifiers of the original acquisition sequence are verified to confirm the continuity and integrity of data acquisition. For example, the sequence is traversed to check the increment and uniformity of timestamps. If a gap is detected, a completion mechanism is triggered. The above verification logic emphasizes the prerequisite of ensuring data integrity, because discontinuous sequences will amplify noise errors in subsequent trend extraction, thereby affecting the decision reliability of the entire oil drilling parameter optimization method.

[0025] If a missing time stamp is found, the original acquisition sequence is completed using an interpolation method to obtain a complete original sequence with a time stamp. The interpolation algorithm is selected based on an adaptive principle of missing duration. For example, a linear interpolation method is used to calculate the data value of the missing point. The linear interpolation logic depends on the linear relationship between the preceding and following data points. That is, the preceding and following values ​​of torque fluctuation amplitude or drilling pressure response delay are used as endpoints, and intermediate values ​​are generated by proportional weighting, such as linear transitions based on time intervals. At the same time, a corresponding time stamp is generated for the missing point to ensure the smoothness and continuity of the completed sequence. The above technical solution, through the multimodal interpolation mechanism, not only restores data integrity but also strengthens the correspondence between various data, such as the linkage smoothing of torque and delay, avoids error amplification in subsequent processing, enhances the robustness of the sequence in drilling in complex formations, and supports the reliable generation of abnormal change pattern stamps.

[0026] Further, in step S1, parameter trend data reflecting formation characteristics are obtained, including: The original sequence is denoised by signal smoothing; high-frequency interference from torque fluctuation amplitude in the original sequence is filtered out by a filtering algorithm; low-frequency variation characteristics related to formation hardness index are retained; and a smoothed parameter trend curve is generated based on the low-frequency variation characteristics. The smoothed parameter trend curve is verified to ensure that it reflects the changing pattern of formation characteristics. If the parameter trend curve has abnormal fluctuations, it is further optimized by secondary smoothing to obtain the final parameter trend data that reflects the formation characteristics.

[0027] Specifically, in one embodiment, the original sequence is processed to filter out interference features and obtain parameter trend data reflecting formation characteristics. This involves first employing signal smoothing techniques to reduce noise in the original sequence. Specifically, the initial dynamic parameter sequence constructed from torque fluctuation amplitude and drill pressure response delay data collected by the sensor array is initially smoothed. The smoothing process is achieved through moving averages or Gaussian filtering, gradually attenuating random noise in the sequence, such as instantaneous disturbances induced by sensor jitter or environmental vibrations, while preserving the core temporal structure of the sequence. This reduces noise interference with subsequent formation trend extraction and ensures that the alignment between torque fluctuation amplitude data and drill pressure response delay data on the time axis is not disrupted. For example, in the complex formation environment of oil drilling, this initial smoothing not only smooths out peak fluctuations in torque data but also strengthens the correlation between delay data and torque data, avoiding misjudgments of false formation hardness changes caused by noise amplification and improving data reliability.

[0028] Furthermore, a filtering algorithm is used to remove high-frequency interference from the torque fluctuation amplitude in the original sequence. This filtering process employs a low-pass filter to target the torque fluctuation amplitude data. The filter's cutoff frequency is set below the typical frequency threshold related to formation hardness index changes, thus isolating the high-frequency components. These high-frequency interferences typically originate from drilling equipment rotor imbalance or downhole transient turbulence, while low-frequency components carry information about gradual changes in formation hardness. The filtering algorithm uses a closed-loop process of frequency domain transformation, such as Fast Fourier Transform, and time domain reconstruction. First, the torque fluctuation amplitude data sequence is mapped to the spectral domain, eliminating frequency band energy above the threshold. Then, it is inversely transformed back to the time domain, ensuring that the filtered torque fluctuation amplitude data maintains a consistent correspondence with the low-frequency portion of the original sequence. Simultaneously, drilling pressure response delay data is used as an auxiliary channel for synchronous filtering, forming spectral linkage between multiple data points. For example, in shallow sandstone and oil drilling scenarios, this filtering helps highlight the slow increasing trend of torque fluctuations, directly corresponding to the transition points of the formation hardness index, thus providing a noise-free basis for parameter optimization and avoiding the implicit accumulation of equipment overload risks.

[0029] Based on this, low-frequency variation characteristics related to formation hardness index are retained. After filtering out high-frequency interference, spectral analysis is performed on the low-frequency band corresponding to the remaining torque fluctuation amplitude data. The changes in the aforementioned low-frequency band usually reflect the gradual change in formation hardness, such as the transition from soft soil to rock. The spectral analysis extracts the amplitude and phase information of low-frequency variation characteristics through peak detection and energy distribution assessment, forming a feature set for subsequent curve generation. The low-frequency amplitude of torque fluctuation directly maps the intensity gradient of formation hardness, while the phase shift of drilling pressure response delay corresponds to the hardness-induced pressure transmission lag adjustment, forming a multi-dimensional vector structure in the feature set. For example, in shale formation oil drilling applications, the retention of low-frequency features can capture the periodic extension trend of delay, reflecting the nonlinear jump of hardness index, thereby improving the reliability of parameter trend data. This retention also strengthens the causal relationship between various data. The method establishes a connection between low-frequency torque amplification and phase delay, ensuring continuous tracking of formation change patterns. Subsequently, based on these low-frequency change characteristics, a smoothed parameter trend curve is generated. The extracted low-frequency feature set is mapped onto the time axis, and a continuous trend curve is generated using spline interpolation to reflect parameter changes under formation influence. The spline interpolation algorithm relies on polynomial piecewise approximation to ensure smooth transitions at key nodes, such as formation transition points. A dual-trajectory correspondence is formed between the low-frequency feature set corresponding to torque fluctuation amplitude data and response delay data. The torque curve captures hardness-related mechanical trends, while the delay curve supplements the temporal sequence of pressure dynamics, thus constructing a unified trend curve framework. For example, in sandstone scenarios during oil drilling, this generation helps accurately mark hardness change points and supports early identification of increased drilling resistance.

[0030] Furthermore, the smoothed parametric trend curves are validated to ensure they reflect the changing patterns of formation characteristics. This involves calculating the local slope and curvature of the trend curves and comparing them with standard patterns in a pre-defined formation change pattern library. For example, linear growth corresponds to uniformly hard formations. If the slope deviation is less than a pre-defined threshold, the validation passes; otherwise, it is marked as a potential anomaly. Subsequently, historical data is compared to confirm whether the curves conform to patterns identified by known formation types. This involves matching local geometric features (slope and curvature) with global patterns and similarity to the historical library, ensuring physical consistency between the slope change of the torque curve and the curvature shift of the delay curve. For example, in complex oil drilling scenarios with fractured limestone, trend validation can involve multi-time-window analysis to check slope stability. If discrepancies are found, an alarm is triggered, ensuring the curves accurately reflect the dynamic changes in formation hardness index. In deep granite or mudstone formations, the above validation is combined with dual-parameter curves to cover highly variable trends such as hardness softening processes, improving system robustness, reducing ineffective adjustments, and supporting the improvement of the overall optimization method's accuracy.

[0031] Finally, if abnormal fluctuations are observed in the aforementioned parameter trend curves, the portion of the fluctuation exceeding a preset upper limit is detected and processed using a secondary low-pass filter or moving average method. The filter parameters are then iteratively adjusted until the fluctuations converge, generating the final parameter trend data. Secondary smoothing is achieved by adjusting adaptive weights; for example, in weighted moving averages, weights are based on historical data comparisons, assigning higher weights and priorities to recent time window data. This ensures that abnormal peaks in the optimized parameter trend curves are smoothed to the formation trend baseline. Fluctuations in the delay curves converge synchronously through linked weights, forming a multi-source fusion correspondence in the final data. For instance, in high-noise gravel-bearing oil drilling environments, this secondary processing makes the trend data more accurately reflect changes in hardness index, avoiding misleading adjustments. In salt rock formations, secondary processing combined with pattern matching supports rotation speed correction and reduces trajectory deviation errors, thereby enriching the diversity of anomaly processing, ensuring applicability in variable formations, and achieving stable optimization of oil drilling parameters.

[0032] Step S2: Based on the parameter trend data, analyze the stratigraphic change pattern and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, then determine the abnormal pattern identifier by comparing with the historical bottom database. In step S2, the corresponding trend feature information is extracted, including: Based on the parameter trend data, a continuous time window is selected for analysis; the trend vector mapping within the continuous time window is calculated; combined with the stratigraphic type identifier, the frequency of potential stratigraphic change patterns in the trend vector mapping is analyzed; based on the frequency of stratigraphic change patterns, trend feature vectors reflecting stratigraphic influence are extracted; the trend feature vectors are standardized, and by analyzing the rate of change of the trend feature vectors, key time points of stratigraphic change are determined, and corresponding trend feature information is generated.

[0033] Specifically, based on the aforementioned parameter trend data, formation change patterns are analyzed, and corresponding trend feature information is extracted. First, based on the aforementioned parameter trend data, a continuous time window is selected for analysis. The length of the time window is dynamically adjusted according to the drilling depth to adapt to the response cycle of different formation environments. For example, in shallow oil drilling scenarios, the time window boundary can be extended to cover the complete drill bit rotation cycle, thereby ensuring the capture of short-term change trends in torque fluctuation amplitude and drilling pressure response delay. The aforementioned dynamic selection mechanism enhances the adaptability of the data processing process, ensures the integrity and continuity of parameter trends within the time window, and achieves precise tracking of formation changes.

[0034] Furthermore, the trend vector mapping within the aforementioned continuous time window is calculated. Specifically, for the torque fluctuation amplitude data within the time window, the least squares method is used to fit the linear trend, obtaining the slope and intercept of the torque trend vector. The least squares method achieves a linear approximation of the torque data by minimizing the sum of squared residuals, ensuring that the slope reflects the overall direction of the gradual change in formation hardness, while the intercept anchors the baseline stability. Subsequently, the drilling pressure response delay data is differentially calculated to generate a delay change vector, highlighting the incremental sensitivity between adjacent points to capture the weak nonlinear effects of pressure transmission. The delay change vector is then compared with the torque trend vector (derived from the slope). The above trend vector mapping is formed by merging the rate and intercept. The merging mechanism here is based on the physical coupling between data, that is, the slope of the torque vector corresponds to the differential amplitude of the delay vector. When the hardness increases, the delay change tends to accelerate, thereby constructing a multi-dimensional mapping structure. For example, in oil drilling operations, the above mapping helps to quantify the direction of parameter changes. For example, the increase in torque indicates an increase in hardness. It supports multi-dimensional expansion in deep well drilling scenarios, such as incorporating the rotation speed dimension to generate a three-dimensional vector. The connection in the algorithm processing is strengthened by aligning the coordinates between vectors, avoiding prediction errors caused by isolated analysis, and ensuring the quantitative continuity of formation influence.

[0035] Based on this, and combined with the formation type identifier, the potential formation change pattern frequencies in the aforementioned trend vector mapping are analyzed. First, the hardness index is extracted from the pre-labeled formation type identifiers. A one-to-one correspondence is established between the vector mapping of torque change and delay change and the classification interval of the hardness index. For example, when sandstone corresponds to a higher hardness interval, the slope gain of the torque vector is preferentially associated. Subsequently, Fourier transform is applied to the trend vector sequence to calculate the frequency spectrum and identify the peak frequencies related to the formation type. The Fourier transform algorithm converts the time domain vector into the frequency domain spectrum. The peak frequencies are highlighted by energy threshold detection to highlight slow change patterns, such as the low-frequency dominance of formation transition. Then, by comparing the frequency distribution under different formation types, the significance of the pattern frequencies is determined and irrelevant noise is filtered out. For example, in the oil drilling transition scenario from sandstone to mudstone, the accelerated change in frequency distribution directly corresponds to the nonlinear jump of the hardness index, thereby revealing the implicit formation transition. The aforementioned pattern frequencies refer to the periodic frequency characteristics of formation changes.

[0036] Subsequently, based on the aforementioned formation change pattern frequencies, trend feature vectors reflecting formation influence are extracted. This involves selecting the amplitude corresponding to the dominant frequency from the frequency spectrum corresponding to the pattern frequencies as feature components. The selection of the dominant frequency is based on peak energy ranking to ensure that the components capture the core oscillation patterns of formation hardness. Next, combining torque fluctuation amplitude data and drilling pressure data, a trend feature vector is constructed. Frequency components, torque amplitude gain, and delay phase difference are used as elements to form a high-dimensional vector structure. The intensity component corresponding to the torque fluctuation amplitude data reflects cutting resistance, and the time delay component corresponding to the delay data reflects transmission efficiency. The vector weights are adjusted using a formation hardness index. For example, formations with high hardness are given higher weights to the amplitude component to enhance the quantification of the true impact. This weighting algorithm emphasizes the role of the index as a regulating factor, ensuring that the proportional relationship between vector elements conforms to physical coupling, such as the orthogonal relationship between amplitude and phase difference. For instance, in high-temperature, high-pressure oil drilling wells, environmental corrections can be added to improve robustness. In complex geological areas, the multifaceted support of amplitude and phase difference components comprehensively describes the formation influence, supporting the continuity of drilling decisions. The extracted vectors are then directly input into the standardization process to avoid feature silos.

[0037] Furthermore, the aforementioned trend feature vectors are standardized to ensure that the numerical range of the trend feature vectors is consistent. The standardization adopts the min-max normalization method, which scales the vector elements to a uniform range. The normalization algorithm logic achieves comparability across time windows through linear transformation, i.e., based on global extrema, ensuring that the torque-dominated amplitude element and the delay-dominated phase element are aligned in scale, while preserving the gradient of relative differences such as hardness-related increases. For example, after standardization, the original vectors of different time windows facilitate a unified benchmark for subsequent rate of change calculation, thereby strengthening the relationship between multiple data and multiple processed data. For example, the standardized amplitude component directly maps to the optimized hardness trend, avoiding mode mismatch caused by scale bias, and supporting the cross-scenario applicability of the overall oil drilling parameter optimization method.

[0038] Finally, by analyzing the rate of change of the elements in the aforementioned trend feature vector, key time points of formation change are determined, and corresponding trend feature information is generated. Specifically, the above analysis obtains the rate of change by calculating the difference between adjacent time window vectors. The differential logic emphasizes incremental sensitivity to quantify the dynamic evolution of vector elements, such as the rate of change of torque fluctuation amplitude corresponding to hardness acceleration. Subsequently, a rate of change threshold is set. When the rate of change exceeds the threshold, it is marked as a key time point. The key time points and corresponding vectors are summarized to generate trend feature information, including timestamps and change descriptions. Here, the summarization mechanism forms a descriptive output chain. For example, when the rate of change exceeds the threshold, a semantic label of "hardness increase transition" is generated and matched with a preset pattern library to verify consistency. For example, in long-term oil drilling data, the rate of change analysis can be performed in segments to cover depth gradients, supporting the detection of gradual changes with low thresholds in the first segment and the capture of abrupt changes with high thresholds in the second segment, thereby accurately locating formation transition points. For example, after the key time point is marked, the parameter adjustment plan is directly triggered to avoid accident risks. The generated information is used as output to connect to the abnormal pattern recognition link to ensure the integrity of the logical closed loop from trend to decision and the accuracy of dynamic adjustment in the overall method.

[0039] Further, in step S2, if the trend feature information meets preset conditions, an abnormal pattern identifier is determined by comparing historical data, including: Calculate the absolute value of the slope of the trend feature information; if the absolute value of the slope exceeds the upper limit of the preset adjustment value, it is determined that the trend feature information meets the preset conditions; extract the expected adjustment direction matching the trend feature information from the pre-established stratigraphic model library by using historical data comparison method; determine the corresponding abnormal change pattern identifier according to the expected adjustment direction; and verify the matching degree of the abnormal change pattern identifier.

[0040] Specifically, in one embodiment, if the aforementioned trend feature information meets preset conditions, an abnormal pattern identifier is determined by comparing historical data. Specifically, the absolute value of the slope of the aforementioned trend feature information is calculated for the smoothed parameter trend curve. The trend vector is processed within a continuous time window, where the slope value is obtained by fitting the curve segment using the least squares method to ensure that the drastic degree of formation change is captured. Here, the algorithm logic of the least squares method emphasizes the joint fitting of torque fluctuation amplitude and drilling pressure response delay data, that is, using the amplitude change dominated by torque as the main independent variable and the delay as an auxiliary weight to minimize the residual, thereby generating a slope index that reflects the jump in formation hardness index. For example, in the continuous time window analysis of oil drilling, the above calculation strengthens the correspondence between the amplitude component and the phase component in the trend feature vector, avoids the deviation caused by a single parameter, and uses the absolute value of the slope as a bridge for quantitative threshold judgment to support the spatiotemporal consistency of subsequent condition judgment.

[0041] Furthermore, if the absolute value of the slope exceeds the preset upper limit of the adjustment value, the trend characteristic information is determined to meet the preset conditions. These preset conditions are based on the upper limit range set by historical drilling data as a dynamic threshold benchmark. When the calculated absolute value exceeds this range, anomaly handling is triggered. This threshold setting originates from the statistical distribution of the formation hardness index, ensuring that the excess slope directly corresponds to potential formation change risks. For example, in soft rock formation oil drilling scenarios, a high slope value indicates an abnormal linkage between torque increase and delay shortening, thereby activating the historical comparison mechanism. The above determination establishes a preliminary causal chain between trend characteristic information and abnormal change patterns through the comparison relationship between the slope and the upper limit range, avoiding false triggering by low thresholds or omission by high thresholds. Furthermore, it also... By comparing historical data, the expected adjustment direction matching the aforementioned trend feature information is extracted from a pre-established stratigraphic model library. This involves first comparing the current trend feature vector sequence with historical sequences (historical trend feature vector sequences) stored in the model library under various stratigraphic hardness indices. Euclidean distance is used to calculate similarity; the algorithm quantifies the geometric proximity between vectors using multidimensional Euclidean norms, ensuring that the amplitude similarity of the torque sequence and the time lag similarity of the delay sequence form a joint score. Subsequently, the adjustment direction, such as adjusting rotation speed and correcting deflection angle, is extracted based on the top few matches with the highest similarity. This extraction is achieved through weighted fusion, where the weights decrease according to similarity, strengthening the correspondence between various historical sequences and the trend feature vector sequence.

[0042] Subsequently, based on the expected adjustment direction, the corresponding abnormal change mode identifier is determined. Specifically, the adjustment direction is mapped to a predefined identifier system. For example, the positive correction direction is mapped to the "hard layer mutation" identifier, which facilitates subsequent parameter linkage. The mapping logic here is based on the semantic dictionary of direction-identifier, ensuring that the torque-dominated acceleration direction and the delay-dominated skew angle correction direction form a multimodal correspondence, thereby generating standardized identifier labels. Furthermore, the above-mentioned abnormal change mode identifiers are classified and stored for subsequent retrieval. The classification is based on the stratum type, such as hard rock or sand layer.

[0043] Finally, the matching degree of the above-mentioned abnormal change pattern identifiers is verified to ensure their accuracy. This includes calculating the cosine similarity between the identifier and the real-time acquired sequence. If the similarity is lower than a preset threshold, the pattern library is re-compared to ensure that the joint projection of the torque sequence and the delay sequence matches the expected pattern of the identifier. Subsequently, the identifier accuracy is adjusted based on the verification results, such as by iteratively updating the weights in the library to improve accuracy. Here, the iterative logic forms a feedback chain. For example, in hard rock oil drilling, the initial identifier is directly used for parameter optimization after verification, while low similarity scenarios trigger re-extraction to ensure the reliability of the identifier and support the stability of continuous drilling operations. This not only overcomes the limitations of a single comparison but also strengthens the robustness of the abnormal identifiers in the overall method through the dynamic connection between similarity and the real-time sequence. Finally, the verified identifiers are output to the initial adjustment direction generation stage to achieve end-to-end accuracy and safety of oil drilling parameter optimization.

[0044] Step S3: For the abnormal mode identifier, generate the parameter adjustment direction for drilling control; optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least rotation speed and motion trajectory deflection angle; In step S3, a preliminary parameter adjustment direction is generated for the abnormal mode identifier, including: Based on the abnormal pattern identifier, a geological change description with high pattern matching accuracy is obtained; the geological change description is analyzed by combining the optimization target priority of rotation speed correction and motion trajectory deflection angle; based on the analysis results, a preliminary parameter adjustment direction sequence is generated; the feasibility of the parameter adjustment direction sequence is verified, and if the parameter adjustment direction sequence exceeds the working range, the preliminary parameter adjustment direction sequence is corrected through constraint conditions to obtain an adjustment direction that meets the requirements.

[0045] Specifically, for the aforementioned anomaly pattern identifiers, preliminary parameter adjustment directions are generated. Based on these anomaly pattern identifiers, descriptions matching the identifiers are retrieved from a pre-established formation pattern library. This ensures that the descriptions capture the formation physical mechanisms implied by the anomaly identifiers. For example, when an anomaly pattern identifier indicates a sudden increase in formation hardness, this retrieval logic prioritizes extracting descriptive entries that indicate an increase in torque fluctuation amplitude caused by hard rock layers. This strengthens the semantic correspondence between anomaly identifiers and formation change descriptions. That is, the amplitude gain dominated by torque fluctuation data directly maps to the description of hardness jumps, while drilling pressure response delay data supplements the auxiliary description of time lag adjustment, forming a descriptive basis for multi-source data fusion. This not only avoids the amplification of biases based on low-precision descriptions but also provides spatiotemporally consistent input support for subsequent analysis. For example, in the hard rock layer scenario of oil drilling, this step ensures that the description reflects the torque overload risk, supporting the continuous logical chain from anomaly detection to parameter optimization in the overall method.

[0046] Furthermore, combining the optimization priorities of rotation speed correction and trajectory deviation angle, the above-mentioned formation change description is analyzed. First, the priority of rotation speed correction is determined to be higher than that of trajectory deviation angle, because rotation speed directly affects drill bit efficiency and torque stability, while deviation angle adjustment focuses more on long-term control of trajectory deviation. Here, the priority logic is based on the weight allocation of the formation hardness index. That is, hardness-related indicators such as torque increase thresholds are first extracted from the formation change description, and the expected correction value of rotation speed is calculated. The calculation uses a proportional mapping algorithm to convert the hardness jump amplitude in the description into a corresponding speed reduction ratio, ensuring a positive correlation between the correction value and torque data. That is, when hardness increases, the speed is moderately reduced to alleviate overload. Subsequently, the deviation angle optimization objective is integrated to evaluate the trajectory. Deviation risk is assessed by comparing the difference between historical trajectory sequences and current delay data. If the deviation exceeds a preset difference, an additional skew angle adjustment is applied after speed correction. Through the coupling relationship between parameters, i.e., rotation speed correction as the dominant variable influencing the subordinate adjustment of skew angle, a dynamic alignment relationship between speed and skew angle is formed. For example, in oil drilling scenarios where soft soil and hard rock layers switch, the assessment can be extended to multiple time windows. The priority of rotation speed remains unchanged, but skew angle optimization is dynamically weighted based on historical trajectory data, thereby supporting the accuracy of the overall adjustment direction. The above priority mechanism improves the response speed and safety of drilling operations. When adjusting the above parameters, other drilling parameters are also adaptively adjusted according to the actual drilling conditions to avoid the implicit accumulation of equipment damage risks.

[0047] Based on the analysis results, a preliminary parameter adjustment direction sequence is generated. The above generation process integrates the velocity correction value and the skew angle adjustment amplitude into a timestamp sequence. The sequence construction logic depends on the output vector of the descriptive analysis. That is, using the time window as an index, the torque-dominated velocity sequence and the delay-dominated skew angle sequence are arranged in parallel to ensure the spatiotemporal synchronization of the parameter direction at each time point. For example, based on the hardness jump description, a gradual sequence of velocity is generated with a moderate decrease from the current value, while the skew angle sequence smoothly transitions to the deviation compensation value. The above parallel arrangement strengthens the relationship between multiple data and multiple processed data. That is, the original torque data corresponds to the amplitude adjustment of the velocity sequence after priority analysis, while the delay data corresponds to the phase shift of the skew angle sequence, thus forming a complete preliminary sequence framework, supporting the input integrity of subsequent verification, and ensuring that the sequence captures the dynamic response of formation transition in the complex formation environment of oil drilling.

[0048] Subsequently, the feasibility of the preliminary parameter adjustment sequence was verified to ensure it conformed to the working range of the drilling equipment. Specifically, the working range data of the drilling equipment was first obtained, including the upper and lower limits of rotation speed and deflection angle. The range data was derived from the joint modeling of equipment specifications and environmental loads. Next, the sequence values ​​were compared item by item. The speed and deflection angle values ​​at each time point were compared with the corresponding ranges for threshold judgment. If the values ​​were within the range, they were marked as passed; otherwise, they were marked as abnormal. The comparison logic here emphasizes the continuity of verification at each time point to ensure the overall physical executability of the sequence. Further, the feasibility score of the overall sequence was calculated. This calculation integrates the pass rate and deviation range through weighted summation to form a score evaluation, where the weights are based on priority. The speed is weighted higher than the deflection angle. If the score exceeds the preset threshold, it is considered compliant. The verification process here particularly emphasizes the integration of external environmental factors. For example, the wind load coefficient is used as a stability assessment indicator. The wind pressure value is calculated using aerodynamic formulas, which is based on the product of the square of the wind speed and the load coefficient. This ensures that the sequence value does not exceed the equipment stress limit caused by wind pressure, thereby strengthening the linkage verification relationship between the speed sequence and the deflection angle sequence. For example, in the low-hardness scenario of shallow oil drilling, this verification covers the comparison of small speed adjustments, while in the high-hardness scenario, it verifies the compliance of the lower limit of large adjustment amplitudes and links with historical feedback to additionally check trajectory stability. This supports multi-faceted safety protection in the overall method and avoids equipment failure caused by exceeding the range.

[0049] Finally, if the initial parameter adjustment direction sequence exceeds the working range, it is corrected by constraints to obtain a suitable adjustment direction. This involves first identifying outliers, such as velocity sequences below the lower limit or skew angles above the upper limit. Then, constraints such as linear interpolation are applied to correct these outliers to the range boundary, while adjacent values ​​are adjusted to maintain sequence smoothness. The interpolation algorithm here relies on the linear transition relationship between preceding and following points, using boundary values ​​as anchors and generating intermediate correction values ​​through time-weighted scaling. This ensures that the torque-dominated velocity sequence and the delay-dominated skew angle sequence maintain coupling consistency after correction. Finally, the corrected sequence is verified to conform to the range, generating the final adjustment direction. Verification involves repeating the aforementioned fractional calculations to form a closed-loop feedback. The correction process is linked to the verification step to ensure no new outliers are introduced, achieving end-to-end stability and accuracy of the overall oil drilling parameter optimization method.

[0050] Further, in step S3, the final optimization guidance data is obtained, including: The parameter adjustment direction sequence is analyzed by parameter linkage weight analysis; the adjustment value time window is dynamically divided by combining the real-time feedback cycle; the parameter adjustment weight in each time window is calculated according to the dynamically divided time window; the preliminary parameter adjustment direction sequence is optimized and adjusted according to the parameter adjustment weight; the final drilling parameter optimization guidance sequence is generated, and the stability of the final drilling parameter optimization guidance sequence is verified.

[0051] Specifically, in one embodiment, the adjustment direction of the above parameters is optimized through weight analysis and real-time feedback to obtain the final optimization guidance data. Specifically, the parameter adjustment direction sequence is analyzed through parameter linkage weight analysis, mainly to identify the linkage relationship between each parameter in the sequence, such as the mutual influence between torque fluctuation amplitude and drilling pressure response delay. The identification logic is based on correlation calculation to ensure that the amplitude change dominated by torque data and the time delay effect dominated by delay data form a coupled weight allocation. That is, the torque increase often amplifies the feedback sensitivity of delay, thereby assigning corresponding weight values ​​to quantify the priority of each direction in the sequence. Here, the weight calculation adopts the statistical covariance method to convert the linkage relationship into numerical coefficients, ensuring that highly correlated parameters (such as the positive dependence of torque and delay) receive higher weights, forming a preliminary priority framework for the sequence. For example, in the hard rock formation scenario of oil drilling, it can quickly assess the dependence between parameters, improve optimization efficiency, and strengthen the relationship between multiple data and multiple processed data. That is, the weight of the torque sequence is directly mapped to the amplification factor of speed correction, while the delay sequence corresponds to the suppression factor of deflection angle adjustment, supporting the continuous logical chain from abnormal response to guidance generation in the overall method.

[0052] Furthermore, the adjustment time window is dynamically divided based on the real-time feedback cycle. This division first collects real-time feedback cycle data, including instant updates of torque fluctuation amplitude and drill pressure response delay provided by the sensor array. The acquisition logic emphasizes the temporal synchronization of the cycle to ensure that the updates of torque data and delay data are aligned to reflect the real-time variation of the formation hardness index. Subsequently, the size of the time window is dynamically adjusted according to the length and frequency of change of the feedback cycle. For example, the time window is reduced when the feedback cycle is shortened to capture rapid changes, and expanded when the cycle is extended to smooth low-frequency trends. Here, the adjustment algorithm is based on an adaptive threshold mechanism, using the torque change frequency as the dominant variable and the delay variation as the secondary variable. As an auxiliary adjustment, a dynamic boundary function is used to form the time window boundary, ensuring that the divided time window adapts to the variability of the formation hardness index. For example, in soft soil oil drilling scenarios, extending the time window helps reduce unnecessary adjustment frequency, thereby improving efficiency and reducing energy consumption. In high-hardness formations in deep well drilling, prioritizing short-cycle feedback generates a more stable sequence, avoiding deviations caused by delayed responses. This dynamic division is not only based on historical data comparison (such as window scaling under threshold triggering), but also enhances the adaptability of the data processing process, ensuring that the torque-dominated short-term window and the delay-dominated long-term window complement each other, supporting the spatiotemporal consistency of subsequent weight calculations, such as... Figure 2 As shown, different time windows are set according to different geological conditions.

[0053] Based on this, according to the dynamically divided time windows mentioned above, the parameter adjustment weights within each time window are calculated. For each time window, a trend vector mapping is extracted from the smoothed parameter trend curve. For example, the slope of torque fluctuation amplitude and the average value of drill pressure response delay are calculated. The extraction logic generates vectors through least squares fitting to ensure that the slope reflects the overall trend of formation changes and the average value captures the stable baseline of the delay. Subsequently, the weights are calculated, for example... The correlation coefficient is the correlation coefficient between torque fluctuation amplitude and delay data, forming the final weight value. The weighted average algorithm emphasizes the contribution weighting of statistical summation, that is, the weight of each parameter is fused according to its importance in the sequence (e.g., the direct impact of delay on safety is higher than the indirect contribution of torque). The weights of all parameters within the time window are summarized to form an adjusted weight set. For example, in the mud layer oil drilling environment, the formation influence can be accurately quantified to ensure that the weight set reflects the actual dynamics. By multiplying the slope-ratio and the correlation coefficient, the connection between torque and delay is strengthened, avoiding the accuracy loss under the static method and supporting the quantitative continuity of formation influence in the overall method.

[0054] Subsequently, the weights are adjusted according to the above parameters, and the sequence of parameter adjustment directions is optimized. Specifically, the calculated weights are first applied to each direction in the sequence for weighted correction. For example, the adjustment magnitude of high-weight directions is amplified, while low-weight directions are moderately suppressed. The correction logic is based on scalar multiplication of the weight vector, ensuring that the priority of rotation speed correction is higher than that of the subordinate adjustment of the trajectory angle. Further, the sequence is iteratively adjusted until the optimization target priority is met. This iteration adopts a gradient descent mechanism to gradually minimize the weight deviation function until the sequence converges to the real-time requirements. Here, the iterative process forms a feedback chain, that is, the torque-dominated speed sequence, after being amplified by high weights, is linked and smoothed with the delay-dominated angle sequence, ensuring that the optimized sequence is more in line with the dynamic environment of oil drilling. For example, in hard rock formations, rotation speed is prioritized to reduce equipment wear. The above adjustments not only strengthen the correspondence between various processed data (such as the weight set directly mapping the magnitude gradient of the sequence), but also support the robustness of the generation process through the priority satisfaction threshold judgment.

[0055] Furthermore, the optimized direction sequence is integrated into a complete guidance sequence, including timestamps and adjustment instructions. The integration logic relies on the parallel splicing of time windows to ensure the spatiotemporal alignment of speed and deflection instructions at each time point, forming an executable sequence framework; such as Figure 3 As shown, the stability of the final drilling parameter optimization guidance sequence is verified to ensure the consistency of the final optimized guidance data across different time windows. This involves first comparing the trend vectors of the guidance sequences within different time windows, such as calculating the variance of the sequence slope. If the variance is below a preset threshold, it is considered consistent. The comparison logic evaluates geometric similarity through vector inner product, ensuring physical coupling between the slope trend of the velocity sequence and the curvature shift of the skew angle sequence. Subsequently, for abnormal windows, historical data is used to compare and correct the sequence. This correction is based on similarity matching from a pattern library, iteratively updating the abnormal vector to ensure overall stability. Finally, the final sequence is output after verification for dynamic adjustment instructions in drilling operations. The verification process emphasizes the closed-loop effect of historical comparison, such as extracting similar patterns from a formation pattern library to confirm matching accuracy, thereby maintaining the stability of the guidance data during the transition between hard rock and soft soil layers, avoiding interruptions caused by sudden adjustments, and covering multi-window evaluation in long-cycle oil drilling to ensure consistency and reduce deviation risks, supporting end-to-end accuracy and safety from optimization to execution in the overall method.

[0056] Step S4: Based on the final optimization guidance data, map it to the control interface and generate dynamic adjustment instructions; specifically including: The final optimized guidance data is input into the navigation control interface; parameter mapping is performed for rotation speed correction and motion trajectory deviation angle; a real-time parameter fine-tuning scheme is generated based on the parameter mapping results; the dynamic adjustment instructions for drilling operations are determined through the real-time parameter fine-tuning scheme; the dynamic adjustment instructions are prioritized, with adjustments to key parameters being executed first. Specifically, in one embodiment, the final drilling parameter optimization guidance sequence is transmitted to the navigation control interface. To ensure that the rotation speed correction value and motion trajectory deflection optimization target at each timestamp in the sequence are completely mapped to the input buffer of the interface, a hash verification algorithm is used for integrity verification. The guidance elements dominated by torque fluctuation amplitude and the guidance elements dominated by drilling pressure response delay are jointly verified to form a spatiotemporal alignment relationship of multi-source data. For example, in the hard rock formation scenario of oil drilling, the synchronization between the optimization sequence and the real-time interface is strengthened to avoid execution deviation caused by transmission delay. The guidance data is used as the basic input to connect to the parameter mapping step, supporting the continuity and reliability of closed-loop control in the overall method.

[0057] Furthermore, the rotational speed correction value and trajectory deflection angle optimization target are extracted from the final optimization guidance data and converted into interface-recognizable mapping parameters. The extraction logic separates the dominant parameter (rotational speed as a direct mapping of torque response) and the subordinate parameter (trajectory deflection angle as an indirect mapping of delay compensation) through a vector decoupling mechanism. This ensures that the difference between the correction value and the current drilling speed is calculated using a linear interpolation method. This linear interpolation algorithm, which inserts new values ​​between known data points, relies on rotational speed data points at adjacent timestamps, generating intermediate correction values ​​through time-weighted averaging to achieve a smooth transition. Subsequently, based on the converted mapping parameters, a joint mapping table for rotational speed and deflection angle is constructed. The correction direction and magnitude within each time window are recorded to form a parameter mapping result. The construction logic here emphasizes the row and column alignment of the table structure, that is, the rows correspond to the time windows and the columns correspond to the parameter types. This ensures that the magnitude correction dominated by torque data directly corresponds to the direction vector in the velocity table, while the time lag effect dominated by delay data corresponds to the magnitude gradient in the skew angle table, forming a coupling matrix of the joint table. For example, in the oil drilling scenario in soft soil, the mapping process combines the historical stratigraphic pattern library to extract the expected adjustment direction to ensure trajectory stability. The multidimensional table structure mentioned above strengthens the connection between various data and various processed data, avoids the accuracy loss caused by isolated mapping, and supports the accurate response of torque smoothing to velocity correction in hard rock formations.

[0058] Based on this, a moving average algorithm is used to average the rotational speed differences in the mapping table corresponding to the parameter mapping results. This moving average process takes the arithmetic mean of continuous subsets of data points. By fusing the differences of the most recent time windows, a smooth fine-tuning value is generated to ensure that the fine-tuning amplitude reflects the gradual trend of the formation hardness index. Subsequently, combined with the real-time feedback cycle, the fine-tuning value is divided into time windows to generate a fine-tuning scheme that includes adjustments to rotational speed and deflection angle. The partitioning algorithm is based on adaptive boundary adjustment of the feedback cycle, using short windows for rapid response to high-frequency torque fluctuations and long windows for stable fusion for trend tracking. The combined mechanism emphasizes the linkage between periodic data and the mapping table. That is, the torque update dominates the amplitude fine-tuning of the short window and the delayed update dominates the phase fine-tuning of the long window, forming a multimodal structure of the scheme. For example, in the oil drilling scenario of high-frequency torque fluctuation, the absolute value of the slope is first extracted from the trend feature vector. If it exceeds the upper limit, the weight of the rotation speed correction is increased, thereby improving the response speed and avoiding the risk of stuck drill. It not only strengthens the logical continuity of data processing steps through the algorithm chain of averaging and partitioning, but also ensures that the fine-tuning scheme captures the dynamic nature of formation change patterns, supporting the adaptability and accuracy of real-time optimization in the overall method.

[0059] Subsequently, the dynamic adjustment instructions for drilling operations are determined through the aforementioned real-time parameter fine-tuning scheme. The conversion logic relies on the semantic mapping of the scheme instructions to ensure that each instruction corresponds to specific parameters of the optimization guidance. For example, speed fine-tuning values ​​are directly converted into speed control codes, and yaw angle fine-tuning values ​​are converted into trajectory deviation compensation codes, forming a serialized set of executable instructions. Here, the sequence construction emphasizes the parallel conversion of time windows, that is, torque-dominated instructions are prioritized for speed adjustment, and delay-dominated instructions supplement yaw angle linkage, ensuring the physical coupling relationship between instructions. For example, in the oil drilling environment where abnormal change pattern indicators appear, the above conversion strengthens the correspondence between the fine-tuning scheme and the operation instructions, supporting the real-time generation of instructions in scenarios with frequent changes in the trend vector mapping of continuous time windows.

[0060] Furthermore, the aforementioned dynamic adjustment commands are prioritized to ensure that adjustments to key parameters are executed first. This involves analyzing the weighted linkages of parameters in the command sequence. The analysis logic calculates the priority score for each command using a weighted summation method. This weighted summation process multiplies the weight of each parameter by its adjustment magnitude and sums the results. The weights are preset based on the formation hardness index (rotation speed correction has a higher weight than deflection angle optimization), forming a total score to quantify priority. Subsequently, the commands are sorted according to their scores, ensuring that high-scoring commands, such as rotation speed adjustments, are executed first. The execution order of the sorted commands is then verified to match the optimization target priority, forming a final sorted list. Verification here uses a similarity matching algorithm, comparing the vector projections of the sorted sequence with the target priority to ensure that torque-dominated high-priority commands and delay-dominated secondary-priority commands form a gradient execution chain. For example, in the soft soil scenario of oil drilling, the sorting process dynamically updates the weights to adapt to changes in formation type, thereby ensuring safe adjustments. This score-sorting-verification algorithm strengthens the dynamic balance between various data in the command sequence, supporting the priority response of commands related to key parameters such as drill pressure response delay, and improving overall drilling efficiency.

[0061] Finally, by monitoring the execution effect of the above-mentioned dynamic adjustment commands in real time, it is confirmed whether the dynamic adjustment commands meet the expected goals. This monitoring collects sensor data after execution and compares it with the expected goals to confirm the degree of compliance. The comparison logic is based on deviation function calculation to ensure that the amplitude of torque execution feedback matches the speed command, and the time delay of delay execution feedback matches the deflection angle command, forming a closed-loop evaluation. For example, in hard rock formation oil drilling, monitoring shows that prioritizing the execution of rotation speed adjustment reduces response delay. The above technical solution, from data input to monitoring confirmation, not only supports the stability of mapping and generation (such as the positive mapping of trend vector slope linked to trajectory optimization), but also ensures the consistency of sorting through multi-faceted expansion (such as dynamic weight adjustment covering diverse scenarios), forming a common argument, namely, improving adjustment efficiency, and seamlessly connecting the confirmed commands to the complete execution chain of oil drilling parameter optimization, achieving end-to-end safety and accuracy.

[0062] This invention also provides an oil drilling parameter optimization system for implementing the above-mentioned method, such as... Figure 4 As shown, the system includes: The data acquisition unit is used to collect data on torque fluctuation amplitude and drilling pressure response delay during the oil drilling process through sensors, and to construct a raw sequence with time stamps. The data preprocessing unit is used to filter and smooth the original sequence to obtain parameter trend data reflecting the formation characteristics; An anomaly identification unit is used to analyze the stratigraphic change pattern based on the parameter trend data and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, the anomaly pattern identifier is determined by comparison with the historical bottom database. The parameter adjustment and optimization unit is used to generate the parameter adjustment direction for drilling control based on the abnormal mode identifier; and to optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least drilling speed and motion trajectory deflection angle. The instruction generation unit is used to map the final optimization guidance data to the control interface and generate dynamic adjustment instructions.

[0063] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the above-described method.

[0064] In summary, this invention, through continuous acquisition and time-series processing of multi-source dynamic parameters during drilling, combined with signal denoising, trend extraction, and feature analysis, can stably and accurately depict the true trend of formation characteristics changing over time, fundamentally improving data quality and the ability to perceive formation changes. By comparing and analyzing with a historical formation pattern database, it can promptly identify abnormal change patterns matching the current working conditions, enabling the system to respond at the initial stage of formation abrupt changes, avoiding the risks associated with traditional reliance on human experience or delayed judgment. For identified abnormal patterns, by introducing parameter optimization target priority and parameter linkage logic, it conducts targeted analysis of formation changes and generates structural data. The system employs a sequence of parameter adjustment directions, ensuring that adjustments to key parameters such as rotation speed and trajectory deflection have a clear basis and inherent order, thus avoiding control instability caused by disordered adjustments of multiple parameters. Simultaneously, by verifying the feasibility and correcting constraints of parameter adjustment directions, the system ensures that all adjustment schemes remain within the safe operating range of the drilling equipment and the working environment, improving the engineering feasibility of the schemes. After the parameter decision-making layer is completed, the system maps the optimized guidance data to the navigation control interface. Through parameter mapping, fine-tuning scheme generation, and command priority sorting, the abstract optimization results are transformed into directly executable dynamic adjustment commands, enabling the drilling equipment to achieve continuous and smooth parameter adjustments according to the rhythm of formation changes. At the same time, by monitoring the execution effect in real time and introducing a feedback mechanism, the adjustment results are used to influence subsequent decision-making processes, forming an adaptive optimization closed loop. Through the synergy of these technical solutions, accurate identification of formation changes, intelligent decision-making and safe execution of drilling parameters are achieved, effectively improving the real-time performance, stability, and safety of parameter adjustments during oil drilling, reducing equipment wear and operational risks, and significantly enhancing the overall efficiency and reliability of drilling operations under complex formation conditions, demonstrating significant engineering application value.

[0065] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0066] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0067] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing oil drilling parameters, characterized in that, The method includes: Step S1: Collect torque fluctuation amplitude and drilling pressure response delay data during the oil drilling process using sensors, and construct a raw sequence with time stamps; perform filtering and smoothing processing on the raw sequence to obtain parameter trend data reflecting formation characteristics; Step S2: Based on the parameter trend data, analyze the stratigraphic change pattern and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, then determine the abnormal pattern identifier by comparing with the historical bottom database. Step S3: For the abnormal mode identifier, generate the parameter adjustment direction for drilling control; optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least rotation speed and motion trajectory deflection angle; Step S4: Based on the final optimization guidance data, map it to the control interface and generate dynamic adjustment instructions.

2. The method as described in claim 1, characterized in that, In step S1, the original sequence with time identifiers is constructed, including: Torque fluctuation amplitude and drill pressure response delay data during oil drilling are acquired in real time using a sensor array. These data are then arranged chronologically to construct an initial dynamic parameter sequence. The initial dynamic parameter sequence is preliminarily labeled based on the formation hardness index, generating a raw acquisition sequence with timestamps. This raw acquisition sequence is stored as input data for subsequent processing, ensuring that each data point in the raw acquisition sequence corresponds to a unique time identifier. The timestamps of the raw acquisition sequence are verified; if any timestamps are missing, they are supplemented using interpolation methods to obtain a complete raw sequence with timestamps.

3. The method as described in claim 2, characterized in that, In step S1, parameter trend data reflecting formation characteristics are obtained, including: The original sequence is denoised by signal smoothing; high-frequency interference from torque fluctuation amplitude in the original sequence is filtered out by a filtering algorithm; low-frequency variation characteristics related to formation hardness index are retained; and a smoothed parameter trend curve is generated based on the low-frequency variation characteristics. The smoothed parameter trend curve is verified to ensure that it reflects the changing pattern of formation characteristics. If the parameter trend curve has abnormal fluctuations, it is further optimized by secondary smoothing to obtain the final parameter trend data that reflects the formation characteristics.

4. The method as described in claim 1, characterized in that, In step S2, the corresponding trend feature information is extracted, including: Based on the parameter trend data, a continuous time window is selected for analysis; the trend vector mapping within the continuous time window is calculated; combined with the stratigraphic type identifier, the frequency of potential stratigraphic change patterns in the trend vector mapping is analyzed; based on the frequency of stratigraphic change patterns, trend feature vectors reflecting stratigraphic influence are extracted; the trend feature vectors are standardized, and by analyzing the rate of change of the trend feature vectors, key time points of stratigraphic change are determined, and corresponding trend feature information is generated.

5. The method as described in claim 4, characterized in that, In step S2, if the trend feature information meets preset conditions, the abnormal pattern identifier is determined by comparing historical data, including: Calculate the absolute value of the slope of the trend feature information; if the absolute value of the slope exceeds the upper limit of the preset adjustment value, it is determined that the trend feature information meets the preset conditions; extract the expected adjustment direction matching the trend feature information from the pre-established stratigraphic model library by using historical data comparison method; determine the corresponding abnormal change pattern identifier according to the expected adjustment direction; and verify the matching degree of the abnormal change pattern identifier.

6. The method as described in claim 1, characterized in that, In step S3, a preliminary parameter adjustment direction is generated based on the abnormal mode identifier, including: Based on the abnormal pattern identifier, a geological change description with high pattern matching accuracy is obtained; the geological change description is analyzed by combining the optimization target priority of rotation speed correction and motion trajectory deflection angle; based on the analysis results, a preliminary parameter adjustment direction sequence is generated; the feasibility of the parameter adjustment direction sequence is verified, and if the parameter adjustment direction sequence exceeds the working range, the parameter adjustment direction sequence is corrected through constraint conditions to obtain an adjustment direction that meets the requirements.

7. The method as described in claim 6, characterized in that, In step S3, the final optimization guidance data is obtained, including: The parameter adjustment direction sequence is analyzed by parameter linkage weight analysis; the adjustment value time window is dynamically divided by combining the real-time feedback cycle; the parameter adjustment weight in each time window is calculated according to the dynamically divided time window; the parameter adjustment direction sequence is optimized and adjusted according to the parameter adjustment weight; the final drilling parameter optimization guidance sequence is generated, and the stability of the final drilling parameter optimization guidance sequence is verified.

8. The method as described in claim 1, characterized in that, In step S4, dynamic adjustment instructions are generated, including: The final optimized guidance data is input into the navigation control interface; parameter mapping is performed for rotation speed correction and motion trajectory deviation angle; a real-time parameter fine-tuning scheme is generated based on the parameter mapping results; dynamic adjustment commands for drilling operations are determined through the real-time parameter fine-tuning scheme; the dynamic adjustment commands are prioritized, with adjustments to key parameters executed first; the execution effect of the dynamic adjustment commands is monitored in real time to confirm whether the dynamic adjustment commands meet the expected goals.

9. An oil drilling parameter optimization system for implementing the method as described in any one of claims 1-8, characterized in that, The system includes: The data acquisition unit is used to collect data on torque fluctuation amplitude and drilling pressure response delay during the oil drilling process through sensors, and to construct a raw sequence with time stamps. The data preprocessing unit is used to filter and smooth the original sequence to obtain parameter trend data reflecting the formation characteristics; An anomaly identification unit is used to analyze the stratigraphic change pattern based on the parameter trend data and extract trend feature information reflecting stratigraphic change; if the absolute value of the slope of the trend feature information meets the preset conditions, the anomaly pattern identifier is determined by comparison with the historical bottom database. The parameter adjustment and optimization unit is used to generate the parameter adjustment direction for drilling control based on the abnormal mode identifier; and to optimize the parameter adjustment direction through parameter weight analysis and real-time feedback to obtain the final optimization guidance data, wherein the parameters include at least rotation speed and motion trajectory deflection angle. The instruction generation unit is used to map the final optimization guidance data to the control interface and generate dynamic adjustment instructions.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the method as described in any one of claims 1-8.

Images