Methods, systems, terminals, and media for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis.

Abstract

This application relates to the field of data analysis technology, and discloses a method, system, terminal, and medium for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis. The method includes: acquiring time-series information and background information of the drilling operation; performing windowing and labeling processing on the time-series information and background information to obtain positive and negative samples; constructing an early warning model and training the early warning model based on the positive and negative samples to obtain a trained early warning model; acquiring current monitoring data during the drilling operation and inputting the current monitoring data into the trained early warning model, outputting early warning results; and providing safety and early warning prompts to drilling personnel based on the early warning results. This application can effectively predict potential faults, thereby improving the accuracy and efficiency of drilling operations and ensuring the safety and stability of drilling operations.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to a method, system, terminal and medium for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis. Background Technology

[0002] In coal mine roadway excavation, drilling is a core component of support and roadway construction, and its efficiency and safety directly impact the excavation progress, project quality, and construction costs. As coal mining extends to deeper and more complex geological conditions, the risk of drilling malfunctions such as stuck drill bits and slippage increases significantly. These malfunctions not only delay construction but can also damage drill bits, reduce roadway stability, and even trigger serious safety accidents.

[0003] Existing drilling equipment generally suffers from insufficient intelligence. Due to the short iteration cycle of drilling rig systems in engineering applications, related construction parameter optimization models and dynamic attitude control algorithms have not yet formed a systematic technical framework. Currently, key parameters such as drilling speed, drilling force, and drill bit rotation speed are mainly set and adjusted based on the operator's experience, lacking quantitative decision-making basis based on theoretical analysis and data-driven approaches. This experience-based control method has significant limitations: under complex geological conditions, the drilling trajectory is prone to nonlinear deviation, manifested as the drill rod axis deviation exceeding the allowable threshold, leading to reduced operational accuracy, shortened equipment life, and increased safety hazards during construction.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main purpose of this application is to provide a method, system, terminal and medium for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis. The aim is to solve the problem in the prior art that the drilling trajectory is prone to nonlinear deviation during drilling operations in complex formation conditions, resulting in low operation accuracy.

[0006] The first aspect of this application provides a method for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis. The method includes the following steps:

[0007] Obtain timing and background information for the drilling operation;

[0008] Based on the time series information and the background information, windowing and labeling processes are performed to obtain positive and negative samples.

[0009] Construct an early warning model, and train the early warning model based on the positive samples and the negative samples to obtain a trained early warning model;

[0010] Acquire current monitoring data during the drilling operation, input the current monitoring data into the trained early warning model, and output the early warning result;

[0011] Based on the aforementioned warning results, safety and early warning prompts will be provided to drilling personnel.

[0012] Optionally, in one embodiment of this application, the timing information is continuous timing data, and the background information includes the total running time and job results;

[0013] The step of performing windowing and labeling processing based on the time series information and the background information to obtain positive and negative samples specifically includes:

[0014] Based on the total running time and the continuous time series data, the window is divided to obtain multiple time windows and their corresponding subsequences, wherein there is some overlapping continuous time series data between two adjacent subsequences.

[0015] Based on the results of the operation, window labeling is performed on the subsequences corresponding to each of the multiple time windows to obtain positive and negative samples.

[0016] Optionally, in one embodiment of this application, the continuous time-series data includes job number, timestamp, drill bit rotation speed data, drilling speed data, and drilling force data, and the multiple sub-sequences include drill bit rotation speed, drilling speed, and drilling force corresponding to each of multiple consecutive time windows;

[0017] The step of dividing the time window based on the total running time and the continuous time series data to obtain multiple subsequences corresponding to each time window specifically includes:

[0018] A time window and step size are set according to each job number, wherein the duration of the time window is greater than the duration of the step size;

[0019] Based on the timestamp, the time window, and the step size, data is extracted from the drill bit rotation speed data, the drilling speed data, and the drilling force data to obtain the drill bit rotation speed, drilling speed, and drilling force corresponding to each of the multiple consecutive time windows.

[0020] Optionally, in one embodiment of this application, the job result includes the final job result corresponding to each time window;

[0021] The step of labeling the subsequences corresponding to multiple time windows based on the operation results to obtain positive and negative samples specifically includes:

[0022] For each time window, if the final result of the operation corresponding to the time window is a fault, and the time between the end time of the time window and the time of the fault occurrence does not exceed the warning threshold, then the time window is determined to be a fault precursor window, and the drill bit rotation speed, the drilling speed, and the drilling force corresponding to the fault precursor window are taken as positive samples; if the time between the end time of the time window and the time of the fault occurrence exceeds the warning threshold, then the time window is determined to be a normal window.

[0023] When the final result of the operation corresponding to the time window is normal, the time window is determined to be a normal window, and the drill bit rotation speed, drilling speed and drilling force corresponding to the normal window are taken as negative samples.

[0024] Optionally, in one embodiment of this application, the positive sample includes a first feature group, and the negative sample includes a second feature group;

[0025] The step of training the early warning model based on the positive and negative samples to obtain a trained early warning model specifically includes:

[0026] Feature extraction is performed on the positive samples and the negative samples respectively to obtain the first feature group corresponding to the positive samples and the second feature group corresponding to the negative samples;

[0027] The early warning model is trained based on the first feature group and the second feature group to obtain a trained early warning model.

[0028] Optionally, in one embodiment of this application, the first feature group includes a first time-domain statistical feature, a first frequency-domain feature, a first trend feature, a first fluctuation feature, a first peak feature, a first cross-correlation feature, and a first cumulative feature, and the second feature group includes a second time-domain statistical feature, a second frequency-domain feature, a second trend feature, a second fluctuation feature, a second peak feature, a second cross-correlation feature, and a second cumulative feature;

[0029] The step of extracting features from the positive samples and the negative samples respectively to obtain a first feature group corresponding to the positive samples and a second feature group corresponding to the negative samples specifically includes:

[0030] High-level feature extraction is performed on the positive samples to obtain the first time-domain statistical features, first frequency-domain features, first trend features, first fluctuation features, first peak features, first cross-correlation features, and first cumulative features corresponding to the positive samples;

[0031] High-level feature extraction is performed on the negative sample to obtain the second time-domain statistical feature, second frequency-domain feature, second trend feature, second fluctuation feature, second peak feature, second cross-correlation feature and second cumulative feature corresponding to the negative sample.

[0032] Optionally, in one embodiment of this application, the step of providing safety and early warning prompts to drilling personnel based on the early warning result specifically includes:

[0033] If the warning result is the fault parameter corresponding to the positive sample, then a fault warning reminder will be given to the drilling personnel;

[0034] If the warning result is a normal parameter corresponding to the negative sample, then the drilling personnel will be given a normal operation prompt.

[0035] A second aspect of this application also provides a drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis, wherein the drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis is applied to the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis described in any of the above solutions; the drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis includes:

[0036] The information acquisition module is used to acquire the timing and background information of the drilling operation;

[0037] The data slicing and labeling module is used to perform windowing and labeling processing based on the time series information and the background information to obtain positive and negative samples;

[0038] The model building and training module is used to build an early warning model and train the early warning model based on the positive samples and the negative samples to obtain a trained early warning model.

[0039] The early warning result output module is used to acquire the current monitoring data during the drilling operation, input the current monitoring data into the trained early warning model, and output the early warning result;

[0040] The early warning module is used to provide safety and early warning prompts to drilling personnel based on the early warning results.

[0041] A third aspect of this application also provides a terminal, wherein the terminal includes: a memory, a processor, and a drilling and anchoring process parameter optimization and fault early warning program based on time-series data analysis stored in the memory and executable on the processor. When the drilling and anchoring process parameter optimization and fault early warning program based on time-series data analysis is executed by the processor, it implements the steps of the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis as described above.

[0042] A fourth aspect of this application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a drilling and anchoring process parameter optimization and fault early warning program based on time-series data analysis, and when the drilling and anchoring process parameter optimization and fault early warning program based on time-series data analysis is executed by a processor, it implements the steps of the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis as described above.

[0043] Beneficial effects: This application provides a method, system, terminal and medium for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis. Based on various time-series sensor data during the drilling process, this application constructs a predictive model for predicting and identifying potential faults, improves prediction accuracy, and achieves effective early warning of potential faults such as stuck drill and slippage, thereby improving the accuracy and efficiency of drilling operations and ensuring the safety and stability of drilling operations. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart of a preferred embodiment of the drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis of this application;

[0046] Figure 2 This is a flowchart illustrating the specific implementation steps of the drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis in a preferred embodiment of this application.

[0047] Figure 3 This is a schematic diagram of data windowing and labeling in a preferred embodiment of the drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis in this application;

[0048] Figure 4 This is a schematic diagram of model training in a preferred embodiment of the drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis in this application;

[0049] Figure 5 This is a structural diagram of a preferred embodiment of the drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis of this application.

[0050] Figure 6 This is a structural diagram of a preferred embodiment of the terminal of this application.

[0051] Explanation of reference numerals in the attached figures:

[0052] 100. Information Acquisition Module; 200. Data Slicing and Labeling Module; 300. Model Building and Training Module; 400. Early Warning Result Output Module; 500. Early Warning Notification Module. Detailed Implementation

[0053] To make the objectives, technical solutions, and effects of this application clearer and more explicit, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. The described embodiments are only possible technical implementations of this application and not all possible implementations. Based on the embodiments in this application, those skilled in the art can obtain other embodiments without creative effort, and these embodiments are also within the protection scope of this application.

[0054] First, the terms used in this application will be explained.

[0055] Industrial data processing: Data generated in industrial scenarios is typically characterized by large scale, high dimensionality, high noise, and heterogeneous origins. Traditional data processing methods mainly rely on manually setting thresholds or rules to filter abnormal signals, but these methods are prone to failure under complex operating conditions. In recent years, data-driven methods have gradually become mainstream. For example, preprocessing techniques based on time-series segmentation, sliding windows, and statistical feature extraction can transform continuous sensor signals into samples suitable for modeling, thereby improving the model's ability to capture transient anomalies and non-stationary patterns. Simultaneously, the introduction of cross-sensor correlation analysis, frequency domain energy decomposition, and cumulative indices provides richer and more robust feature representations for the dynamic monitoring of complex mechanical systems. These methods are widely used in industrial predictive maintenance and equipment health management, but there is still room for improvement in real-time performance and robustness.

[0056] Industrial Parameter Optimization: In typical industrial operations such as drilling and anchoring, machine tool processing, and energy extraction, the setting of process parameters (such as rotational speed, feed rate, and load magnitude) directly affects efficiency and safety. Existing parameter optimization methods mainly rely on empirical rules, offline experiments, or heuristic searches, which suffer from insufficient adaptability and difficulty in handling dynamic changes in operating conditions. In recent years, research combining machine learning and optimization theory has gradually increased. Common methods include using supervised learning models to predict performance under different parameters and then determining the optimal parameter combination through search or optimization algorithms; or using reinforcement learning to achieve dynamic parameter control. These methods can achieve adaptive parameter adjustment to a certain extent, but their effectiveness depends on high-quality data support and is still limited by model computational efficiency and interpretability in practical deployment.

[0057] Extreme Gradient Boosting (XGBoost) is an improved gradient boosting tree algorithm. Its basic principle is to iteratively train multiple weak classifiers (decision trees) and continuously optimize the residuals to improve overall prediction accuracy. Compared to traditional GBDT (Gradient Boosting Decision Tree), XGBoost introduces L1 / L2 regularization terms into the objective function, effectively mitigating overfitting, and significantly improves training efficiency through feature-level parallel computation and a distributed architecture. In engineering applications, XGBoost exhibits strong robustness and scalability, automatically handling missing values, supporting weight adjustments for imbalanced samples, and outputting feature importance rankings to provide explanations for decision-making. Therefore, this algorithm is widely used in structured data modeling scenarios such as financial risk control, medical diagnosis, and industrial monitoring.

[0058] The embodiments of this application can be applied to the following scenarios: intelligent process optimization and fault prediction during drilling and anchoring operations.

[0059] This application fully utilizes multi-source sensor data during drilling to identify and predict potential faults in real time. Specifically, by extracting features and modeling multi-dimensional time-series data during drilling, effective early warnings can be provided before faults occur, thus enabling proactive intervention and scientific guidance of the drilling process. This method not only helps improve the safety and stability of drilling operations and reduce equipment failure rates and maintenance costs, but also provides solid theoretical and data support for the dynamic optimization and intelligent control of construction parameters, which is of great significance for promoting the development of coal mine tunneling equipment towards intelligence and autonomy.

[0060] The core objective of this application is to utilize data from multiple time-series sensors during the drilling process (including drilling speed) to... Drilling force Drill bit rotation speed This research aims to construct an intelligent model capable of predicting and identifying potential faults (including stuck drill bits and slippage), ultimately forming a guidance plan for the drilling process and providing implementation parameter suggestions and fault warnings for practical applications. This prediction task is essentially a time-series binary classification problem with several key unique characteristics: First, the positive and negative samples are extremely imbalanced, with normal drilling samples making up the vast majority and fault samples being scarce but extremely valuable; second, the fault mechanisms are complex, often resulting from the coupling of multiple parameters (such as a decrease in velocity accompanied by drastic fluctuations in force); third, there is a strong need for model interpretability so that engineers can understand the basis for the warnings and verify whether they conform to physical laws; and fourth, the model must ultimately be deployed on-site, requiring high computational performance and good robustness to handle possible sensor noise and data gaps.

[0061] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0062] The preferred embodiment of this application describes a drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis, such as... Figure 1 As shown, the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis includes the following steps:

[0063] In step S101, the timing information and background information of the drilling operation are obtained.

[0064] In one possible implementation, the timing information is continuous timing data, and the background information includes the total running time and job results.

[0065] Specifically, the dataset was constructed and cleaned. The dataset covers key parameters such as operating conditions, drill bit rotation speed, drilling speed, and drilling force, and also includes successful and failed drill-anchoring samples, organized in a table structure. The data within the tables includes time-series data and metadata. Time-series data refers to the data returned by sensors that changes over time (e.g.,...). Figure 2 As shown, the drill bit rotation speed Drilling speed Drilling force As shown in Table 1, the background information refers to the current drilling conditions (layer type, drill type, outcome label), as shown in Table 2. It can be understood that during data preparation, the raw monitoring data (i.e., time-series information) and related metadata (i.e., background information) of the drilling operation are acquired. The monitoring data includes the operation number (run_id), timestamp, and drilling speed. Drilling force Rotation speed Time-series variables; background information includes operating conditions, total running time (run_time), and final result (outcome in Table 2: 1 indicates a fault, 0 indicates normal).

[0066] Table 1: Temporal Information Section of the Dataset

[0067]

[0068] Table 2: Background Information Section of Data Set

[0069]

[0070] In step S102, windowing and labeling are performed based on the time sequence information and the background information to obtain positive and negative samples.

[0071] In one possible implementation, windows are divided according to the total running time and the continuous time series data to obtain multiple time windows corresponding to their respective subsequences; wherein, there is some overlapping continuous time series data between two adjacent subsequences; according to the job results, the subsequences corresponding to the multiple time windows are labeled to obtain positive samples and negative samples.

[0072] It should be noted that, see Figure 2 and Figure 3 Windowing and labeling are performed to transform drilling and anchor time-series data into a dataset suitable for supervised learning. This application employs a sliding window-based segmentation and labeling strategy. This method is a key data preprocessing step for achieving accurate identification of fault precursors.

[0073] Windowing is the extraction of engineering-significant local segments from a complete time series. Specifically, a fixed-length time window (e.g., 10 seconds) is set, and the time is slid along the time axis in steps much smaller than the window length (e.g., 1 second), thereby generating a large number of overlapping continuous subsequences. This overlapping sampling method ensures that any critical moment in the drilling process is covered, effectively avoiding the risk of critical signals being truncated by the window boundaries, thus providing a data foundation for capturing transient or periodic fault precursor features.

[0074] Labeling is the process of assigning a semantic label to each window sample that can represent its future state. Specifically, a warning label definition rule based on a future time threshold is adopted: for a window with an end time of t_end, if the drilling operation (run_id) to which it belongs eventually fails, and satisfies 0 ≤ (run time - t_end) ≤ T_warn (where T_warn is a preset warning time window, such as 5 seconds), then the window is marked as a positive sample (1), indicating that the segment has contained sufficient information about the precursors of the failure; otherwise, it is marked as a negative sample (0).

[0075] This rule strictly ensures the forward-looking nature of the labels, meaning that the model's learning objective is to predict whether a fault will occur in the near future based on the current window information, thus meeting the practical needs of engineering practice for early warning systems to provide sufficient response time.

[0076] In one possible implementation, the continuous time-series data includes job number, timestamp, drill bit rotation speed data, drilling speed data, and drilling force data. The multiple sub-sequences each include drill bit rotation speed, drilling speed, and drilling force corresponding to multiple consecutive time windows. A time window and step size are set according to each job number, wherein the duration of the time window is greater than the duration of the step size. Data is extracted from the drill bit rotation speed data, drilling speed data, and drilling force data based on the timestamp, the time window, and the step size to obtain the drill bit rotation speed, drilling speed, and drilling force corresponding to each of the multiple consecutive time windows.

[0077] First, let's explain the following parameters: Window length W: The value ranges from 5 to 15 seconds. This range is chosen based on the timescale at which fault precursors typically occur, combined with the acceptable response time in actual engineering. Step size S: 1 to 2 seconds (sliding window), representing the interval at which the sliding window moves along the time axis. When the sequence length is too short, a single-window partitioning method is used, setting the step size to S=W, retaining only the data from the first window to maximize sample utilization efficiency and avoid data fragmentation. Warning window T_warn: Used to limit the look-ahead range of the model's prediction, i.e., determining how long in the future a fault is likely to occur. Its specific value depends on the response time allowed by on-site operations and is usually determined by engineering safety specifications or actual application requirements.

[0078] Specifically, during the window partitioning process, for each job number run_id, the window length W and step size S are set. From the job start time to the maximum timestamp t_max, a continuous time window [t_start, t_end] is extracted in a sliding manner, and the corresponding time series data within the window is extracted to form subsequences (multiple).

[0079] In one possible implementation, the operation result includes the final operation result corresponding to each of the time windows. For each time window, if the final operation result corresponding to the time window is a fault, and the time between the end time of the time window and the time of the fault occurrence does not exceed a warning threshold, then the time window is determined to be a fault precursor window, and the drill bit rotation speed, drilling speed, and drilling force corresponding to the fault precursor window are taken as positive samples; if the time between the end time of the time window and the time of the fault occurrence exceeds a warning threshold, then the time window is determined to be a normal window; if the final operation result corresponding to the time window is normal, then the time window is determined to be a normal window, and the drill bit rotation speed, drilling speed, and drilling force corresponding to the normal window are taken as negative samples.

[0080] Specifically, during the window labeling process, each time window is labeled. If the final result of the corresponding job is outcome = 1, and the window end time t_end is less than the fault occurrence time run_time, the warning threshold T_warn is not exceeded, then the window is determined to be a "fault precursor window" and the label is recorded as 1; otherwise, the label is recorded as 0.

[0081] During the sample construction process, the statistical features (such as mean, standard deviation, energy spectrum features, etc.), label value, job number run_id, and window time range [t_start, t_end] corresponding to each window are combined to form a sample.

[0082] Finally, the dataset is generated. The above steps are repeated until all job numbers (run_id) have been processed, resulting in a dataset consisting of multiple window samples, which can be used to train and validate the fault prediction model.

[0083] In step S103, an early warning model is constructed, and the early warning model is trained based on the positive samples and the negative samples to obtain a trained early warning model.

[0084] In one possible implementation, the positive samples include a first feature group, and the negative samples include a second feature group. Feature extraction is performed on the positive and negative samples respectively to obtain the first feature group corresponding to the positive samples and the second feature group corresponding to the negative samples; the early warning model is then trained based on the first and second feature groups to obtain a trained early warning model.

[0085] In one possible implementation, the first feature group includes a first time-domain statistical feature, a first frequency-domain feature, a first trend feature, a first fluctuation feature, a first peak feature, a first cross-correlation feature, and a first cumulative feature; the second feature group includes a second time-domain statistical feature, a second frequency-domain feature, a second trend feature, a second fluctuation feature, a second peak feature, a second cross-correlation feature, and a second cumulative feature. High-level feature extraction is performed on the positive samples to obtain the corresponding first time-domain statistical feature, first frequency-domain feature, first trend feature, first fluctuation feature, first peak feature, first cross-correlation feature, and first cumulative feature; high-level feature extraction is performed on the negative samples to obtain the corresponding second time-domain statistical feature, second frequency-domain feature, second trend feature, second fluctuation feature, second peak feature, second cross-correlation feature, and second cumulative feature.

[0086] Specifically, stage features are extracted from both positive and negative samples. Stage feature extraction is a crucial step in drilling and anchor fault early warning modeling. Its core objective is to transform the windowed raw time-series signal into a high-level feature representation that can characterize the dynamic features of the drilling process. Since sensor signals often exhibit significant fluctuations and non-stationarity during drilling, single-point raw data is insufficient to directly support the model's effective identification of fault precursors. Therefore, this application constructs multi-dimensional statistical and physical features within each time window, covering different levels such as the time domain, frequency domain, and intervariate correlation (see Table 3 for details).

[0087] The following explains the feature groups corresponding to positive and negative samples.

[0088] Time-domain statistical characteristics. Regarding drilling speed: average speed. ; speed standard deviation Median speed Minimum speed Maximum speed ; Interquartile range of velocity Regarding drilling force: average force Standard deviation of force Minimum value of force ; Maximum force ; interquartile range of force Energy level: root mean square of velocity ; root mean square of force It reflects the fluctuations in speed and force. The significance of time-domain statistical characteristics is that typical signal changes often accompany the occurrence of faults, such as a decreasing trend in drilling speed, an increase in drilling force, or a significant increase in the fluctuation amplitude of both.

[0089] Trend or slope characteristics. , "Slope" means "inclination or gradient," and here it indicates the velocity within the window. or drilling force The slope of the linear fit over time; ; ; Represents the final value of the velocity within the window. with initial value difference; Indicates the final value of the drilling force within the window. with initial value The difference. Significance of trend characteristics: Compared with instantaneous values, the continuous upward or downward trend of a variable within a time window is more stable in reflecting potential abnormal operating conditions.

[0090] Fluctuation or mutation characteristics. This is an abbreviation for "Maximum Rolling Standard Deviation of Velocity," indicating the speed within the window. The maximum value of the moving standard deviation; This is an abbreviation for "Maximum Rolling Standard Deviation of Force," representing the drilling force within the window. The maximum value of the moving standard deviation; This is an abbreviation for "Maximum First-Order Difference," representing the absolute value of the maximum drilling speed within the window. or absolute value of drilling force Significance of fluctuation characteristics: Faults such as stuck drills or slippage are often accompanied by increased vibration or sudden signal pulses. The sliding standard deviation and first-order difference can effectively characterize such anomalies.

[0091] Peak or event characteristics. This is an abbreviation for "Number of Peaks in Force," meaning the number of peaks in the force signal that exceed a threshold. It is based on a threshold or... The result of the function; This is an abbreviation for "Maximum Peak Height," which means the maximum height of the peak value, representing the maximum fluctuation amplitude of the signal. This is an abbreviation for "Average Peak Prominence," which refers to the degree to which a peak stands out relative to its surrounding valleys. Significance of peak characteristics: Impact or fracture processes typically exhibit peak enhancement in the force signal; the number, amplitude, and significance of these peaks can serve as important characterizations of the relevant event.

[0092] Frequency domain (FFT / energy band) characteristics. This is an abbreviation for "Total FastFourier Transform Energy," which represents the total energy in the entire frequency domain and reflects the overall signal strength. This is an abbreviation for "Low-band Fast Fourier Transform Energy," which refers to energy in the low-frequency range, i.e., the slowly changing component. This is an abbreviation for "High-band Fast Fourier Transform Energy," which refers to energy in the high-frequency range, i.e., the fast oscillation component. This is an abbreviation for "Dominant Frequency," meaning the dominant frequency with the highest energy, reflecting the main vibration mode. Significance of frequency domain characteristics: Mechanical resonance or friction introduces high-frequency components; the energy distribution and dominant frequency characteristics in different frequency bands help distinguish different types of fault modes.

[0093] Cross-correlation or ratio characteristics. This is an abbreviation for "Correlation between Velocity and Force," which refers to the correlation coefficient between velocity and drilling force, representing the linear relationship between the two. This is an abbreviation for "Lag of Maximum Correlation," which refers to the lag time at which the maximum correlation coefficient is reached, indicating a causal or delayed relationship. This is an abbreviation for "Mean of Force over Velocity," meaning the average force-to-velocity ratio, representing the force per unit velocity. The significance of the cross-correlation characteristic: During normal drilling, a relatively stable coupling relationship is maintained between velocity and force; however, if a fault occurs, this coupling mode and its correlation often change significantly.

[0094] Cumulative or contextual features (cross-window features). This is "Time Since Start of Run," meaning the time elapsed since the start of the task. This refers to "Cumulative Distance," which represents the total drilling progress. The significance of cumulative characteristics: Wear, aging, and cumulative effects during drilling have a significant impact on the occurrence of failures; the cumulative characteristics across windows can reflect these long-term effects.

[0095] Table 3: Feature Extraction Expression Levels and Formulas

[0096]

[0097] Continued from Table 3

[0098]

[0099] Specifically, see Figure 4 Model training and validation are performed. During the model training phase, this application adopts a per-run instance approach. Grouping and segmenting ( The data partitioning strategy strictly separates data from different drilling processes into training and validation sets to effectively avoid data leakage and overfitting issues. This method ensures that validation set samples are completely invisible during training, thus more realistically reflecting the model's generalization ability in real-world application scenarios. For algorithm selection, XGBoost was chosen as the core model. This method not only performs well with small to medium-sized samples but also exhibits good interpretability and robustness. By training on windowed and feature-extracted samples, the model's learning objective is to determine the future performance of the current window. The system detects whether a fault will occur within a time threshold, thereby enabling sensitive capture of key precursor signals. The extracted data features serve as the primary basis for training the classifier in the model.

[0100] Addressing the specific needs of drill and anchor failure early warning tasks, the XGBoost algorithm demonstrates several significant advantages, which are analyzed in detail below: First, its effectiveness in handling the imbalanced sample problem; XGBoost... The parameters (i.e., the core parameters in the gradient boosting tree model for handling imbalanced binary classification samples) explicitly adjust the class weights, increasing the model's attention to rare fault samples. Its regularization mechanism and ensemble learning framework effectively avoid overfitting, controlling the false positive rate while maintaining high recall, perfectly adapting to business scenarios with extreme imbalance between positive and negative samples. Second, its ability to model complex nonlinear relationships and multi-parameter coupling, based on a tree-structured ensemble learning mechanism, enables XGBoost to automatically capture complex interactions between multi-dimensional features (such as the coupling relationship between drilling force fluctuations and speed decreases), eliminating the need for manually constructing feature interaction terms and greatly enhancing the model's ability to express complex fault mechanisms. Third, it provides highly interpretable decision support; XGBoost can output feature importance rankings (such as... (etc.), intuitively displaying key influencing parameters. Combined with interpretive tools such as SHAP, the basis for prediction can be analyzed case by case, enabling engineers to verify whether the model's decisions conform to the physical laws of equipment operation, enhancing system credibility. Fourth, meeting the efficiency and robustness requirements of engineering deployment, XGBoost possesses excellent computational performance, achieving millisecond-level predictions after training, meeting real-time monitoring needs. It is insensitive to missing data and noise, adapting to the uncertainty of on-site data collection, ensuring stable performance under different operating conditions, and possessing strong engineering practical value. During training, the model is continuously iterated and optimized, and evaluation indicators such as accuracy and recall are monitored in real time on the validation set, ensuring that the model maintains a low false alarm rate while capturing sparse fault samples. Through hyperparameter tuning (such as learning rate, tree depth, class weights, etc.), the optimal model configuration is gradually obtained. After model convergence, it not only achieves stable prediction performance on the validation set but also outputs feature importance ranking, providing interpretable decision support for engineers. After the verification phase is completed, the model parameters are solidified and deployed into the early warning system, which has the ability to process sensor data in real time and output risk warnings quickly. This meets the requirements of engineering applications, namely, that the model should provide stable, reliable and interpretable fault precursor identification results while maintaining high inference speed, so as to provide solid support for improving the safety and efficiency of drilling and anchoring operations.

[0101] In step S104, the current monitoring data during the drilling operation is acquired, and the current monitoring data is input into the trained early warning model to output the early warning result.

[0102] Specifically, the current monitoring data includes the current timestamp, current drilling speed, drilling force, and drilling rotation speed. This data is then input into the trained early warning model, which outputs early warning results to provide guidance solutions based on the results.

[0103] In step S105, safety and early warning prompts are given to the drilling personnel based on the early warning results.

[0104] In one possible implementation, if the warning result is a fault parameter corresponding to the positive sample, a fault warning reminder is given to the drilling personnel; if the warning result is a normal parameter corresponding to the negative sample, a normal operation reminder is given to the drilling personnel.

[0105] In this application, by introducing a fault prediction and intelligent decision-making method based on multi-source sensor data, the system optimization and intelligent control of key parameters during coal mine roadway drilling were achieved, yielding the following beneficial effects: First, it is the first time that fault prediction and identification capabilities have been realized, through the control of drilling speed... Drilling force Drill bit rotation speed Modeling with multi-dimensional time-series data enables the prediction and identification of potential faults such as stuck drill bits and slippage, achieving intelligent early warning functions that traditional manual experience cannot cover. Experimental results show that the model's prediction accuracy can reach 92-93%, providing construction personnel with a scientific basis for decision-making and operational intervention. Secondly, it can optimize construction parameters and improve efficiency. Based on the guidance of the intelligent model, drilling parameters were optimized, making the drilling process more stable and efficient: the drilling speed was optimized from an average of 0.026 m / s to an average of 0.029 m / s (an improvement of approximately 11.5%); the drilling force was optimized from an average of 4970 N to an average of 5160 N (an improvement of approximately 3.8%); and the rotational speed was optimized from an average of 608 r / min to an average of 616 r / min (an improvement of approximately 1.3%). This optimized combination mechanism significantly improved the matching state between the drill bit and the rock formation, increasing the overall drilling efficiency by approximately 12%, while reducing drilling downtime and abnormal adjustment time. Third, it improves equipment lifespan and operational stability. The optimized drilling strategy significantly reduces parameter fluctuations: the standard deviation of drilling force is reduced by approximately 38%, and the fluctuation range of rotational speed is reduced by approximately 20%. This stabilization effect effectively slows down fatigue wear on the drill pipe and drill bit, extending the average service life of the equipment by approximately 15% and reducing maintenance frequency by approximately 10%. Fourth, it enhances safety and intelligence. Through the analysis of fault factors output by the interpretable model, engineers can clearly understand the basis for predictions and perform manual verification, ensuring the consistency between the model's prediction results and actual physical laws. Field tests show that after the application of this embodiment, the construction safety accident rate is reduced by approximately 30%, the reliance on manual experience for intelligent parameter control is significantly reduced, and the automation and safety level of coal mine roadway construction is significantly improved.

[0106] In summary, this application is the first to achieve intelligent prediction and identification of potential faults in the drilling process, while also achieving comprehensive improvements in construction efficiency, equipment lifespan, and safety, demonstrating good engineering application value and promising prospects for promotion.

[0107] Next, referring to the accompanying drawings, the drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis proposed in the embodiments of this application is described, and applied to the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis in any of the above schemes.

[0108] Figure 5 This is a structural diagram of the drilling and anchoring process parameter optimization and fault early warning system based on time-series data analysis according to an embodiment of this application.

[0109] like Figure 5 As shown, the drilling and anchoring process parameter optimization and fault early warning system based on time series data analysis includes: an information acquisition module 100, a data slicing and annotation module 200, a model building and training module 300, an early warning result output module 400, and an early warning prompt module 500.

[0110] Specifically, the information acquisition module 100 is used to acquire the timing information and background information of the drilling operation;

[0111] The data slicing and labeling module 200 is used to perform windowing and labeling processing based on the time series information and the background information to obtain positive samples and negative samples;

[0112] The model building and training module 300 is used to build an early warning model and train the early warning model based on the positive samples and the negative samples to obtain a trained early warning model.

[0113] The early warning result output module 400 is used to acquire the current monitoring data during the drilling operation, input the current monitoring data into the trained early warning model, and output the early warning result;

[0114] The early warning module 500 is used to provide safety and early warning prompts to drilling personnel based on the early warning results.

[0115] Figure 6 A structural diagram of a terminal provided in an embodiment of this application. The terminal may include:

[0116] The memory 501, the processor 502, and the computer program stored on the memory 501 and capable of running on the processor 502.

[0117] When the processor 502 executes the program, it implements the drilling and anchoring process parameter optimization and fault early warning method based on time-series data analysis provided in the above embodiments.

[0118] Furthermore, the terminal also includes:

[0119] Communication interface 503 is used for communication between memory 501 and processor 502.

[0120] The memory 501 is used to store computer programs that can run on the processor 502.

[0121] Memory 501 may include high-speed RAM memory, and may also include non-volatile memory. volatile memory), for example, at least one disk storage.

[0122] If the memory 501, processor 502, and communication interface 503 are implemented independently, then the communication interface 503, memory 501, and processor 502 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EIS) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0123] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.

[0124] Processor 502 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of this application.

[0125] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for optimizing drilling and anchoring process parameters and providing early warning of faults based on time-series data analysis.

[0126] One embodiment of this application provides a computer program product, including a computer program that, when executed by a processor, implements the features described in this application. Figure 1 The corresponding embodiments provide a drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis.

[0127] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0128] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0129] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0130] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable storage medium could be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0131] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0132] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0133] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0134] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

[0135] It should be understood that the application of this application is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for drilling and anchoring process parameter optimization and fault early warning based on time series data analysis, characterized in that, The drilling anchor process parameter optimization and fault early warning method based on time series data analysis comprises: obtaining time series information and background information of drilling operation; performing windowing and labeling processing according to the time series information and the background information to obtain positive samples and negative samples; constructing an early warning model and training the early warning model according to the positive samples and the negative samples to obtain a trained early warning model; wherein the type of the early warning model is an XGBoost model; obtaining current monitoring data during drilling operation and inputting the current monitoring data into the trained early warning model to output an early warning result; providing safety and early warning prompts to drilling personnel according to the early warning result; the time series information is continuous time series data, and the background information includes total running time and operation result; the windowing and labeling processing according to the time series information and the background information to obtain positive samples and negative samples specifically comprises: performing window division according to the total running time and the continuous time series data to obtain a plurality of time windows each corresponding to a subsequence, wherein adjacent two subsequences have overlapping parts of continuous time series data; performing window labeling on the subsequences corresponding to the plurality of time windows according to the operation result to obtain positive samples and negative samples; the continuous time series data includes operation number, timestamp, drill bit speed data, drilling speed data and drilling force data, and the plurality of subsequences include drill bit speed, drilling speed and drilling force corresponding to each of the plurality of time windows; the window division according to the total running time and the continuous time series data to obtain a plurality of time windows each corresponding to a subsequence specifically comprises: setting time window and step length according to each operation number, wherein the time window has a length greater than that of the step length; extracting data from the drill bit speed data, the drilling speed data and the drilling force data according to the timestamp, the time window and the step length to obtain drill bit speed, drilling speed and drilling force corresponding to each of the plurality of time windows; the operation result includes operation final result corresponding to each time window; the window labeling on the subsequences corresponding to the plurality of time windows according to the operation result to obtain positive samples and negative samples specifically comprises: for each time window, when the operation final result corresponding to the time window is fault, and the end time of the time window is not more than a warning threshold from the fault occurrence time, the time window is determined as a fault precursor window, and the drill bit speed, the drilling speed and the drilling force corresponding to the fault precursor window are taken as positive samples; if the end time of the time window is more than the warning threshold from the fault occurrence time, the time window is determined as a normal window; when the operation final result corresponding to the time window is normal, the time window is determined as a normal window, and the drill bit speed, the drilling speed and the drilling force corresponding to the normal window are taken as negative samples; the positive samples include a first feature group, and the negative samples include a second feature group; The first feature group includes a first time domain statistical feature, a first frequency domain feature, a first trend feature, a first fluctuation feature, a first peak value feature, a first cross-correlation feature, and a first cumulative feature, and the second feature group includes a second time domain statistical feature, a second frequency domain feature, a second trend feature, a second fluctuation feature, a second peak value feature, a second cross-correlation feature, and a second cumulative feature. 2.The drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis according to claim 1, wherein, The training of the early warning model according to the positive samples and the negative samples comprises: performing feature extraction on the positive samples and the negative samples respectively to obtain a first feature group corresponding to the positive samples and a second feature group corresponding to the negative samples; training the early warning model according to the first feature group and the second feature group to obtain a trained early warning model. 3.The drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis according to claim 2, wherein, The feature extraction on the positive samples and the negative samples respectively to obtain a first feature group corresponding to the positive samples and a second feature group corresponding to the negative samples comprises: performing high-level feature extraction on the positive samples to obtain a first time domain statistical feature, a first frequency domain feature, a first trend feature, a first fluctuation feature, a first peak value feature, a first cross-correlation feature, and a first cumulative feature corresponding to the positive samples; performing high-level feature extraction on the negative samples to obtain a second time domain statistical feature, a second frequency domain feature, a second trend feature, a second fluctuation feature, a second peak value feature, a second cross-correlation feature, and a second cumulative feature corresponding to the negative samples.

4. The method of claim 1-3, wherein, The safety and early warning prompt to the drilling personnel according to the early warning result comprises: if the early warning result is a fault parameter corresponding to the positive sample, performing fault early warning prompt to the drilling personnel; if the early warning result is a normal parameter corresponding to the negative sample, performing normal operation prompt to the drilling personnel.

5. A system for drilling and anchoring process parameter optimization and failure warning based on time series data analysis, characterized in that, The drilling and anchoring process parameter optimization and fault early warning system based on time series data analysis is applied to the drilling and anchoring process parameter optimization and fault early warning method based on time series data analysis according to any one of claims 1-3. The drilling and anchoring process parameter optimization and fault early warning system based on time series data analysis comprises: an information acquisition module configured to acquire time series information and background information of drilling operation; a data slicing and labeling module configured to perform windowing and labeling processing according to the time series information and the background information to obtain positive samples and negative samples; a model construction and training module configured to construct an early warning model and train the early warning model according to the positive samples and the negative samples to obtain a trained early warning model; an early warning result output module configured to acquire current monitoring data in the drilling operation process, input the current monitoring data into the trained early warning model, and output an early warning result; an early warning prompt module configured to perform safety and early warning prompt to the drilling personnel according to the early warning result.

6. A terminal, characterized by comprising: The terminal comprises a memory, a processor, and a time-series data analysis based drilling anchor process parameter optimization and failure warning program stored on the memory and executable on the processor, and the time-series data analysis based drilling anchor process parameter optimization and failure warning program, when executed by the processor, implements the steps of the time-series data analysis based drilling anchor process parameter optimization and failure warning method according to any one of claims 1-3.

7. A computer readable storage medium characterized in that, The computer readable storage medium stores a time-series data analysis based drilling anchor process parameter optimization and failure warning program, and the time-series data analysis based drilling anchor process parameter optimization and failure warning program, when executed by the processor, implements the steps of the time-series data analysis based drilling anchor process parameter optimization and failure warning method according to any one of claims 1-3.

Images