Method for comprehensive quality evaluation of casing running process

Abstract

The application provides a casing running process comprehensive quality evaluation method, comprising the following steps: step 1, collecting casing data and equipment data to be operated and real-time data of the operation site; step 2, inputting the data into a bidirectional time sequence model to monitor the process data in real time and evaluate the data content; step 3, monitoring the hook static time during the casing operation in real time; step 4, monitoring the lowering speed during the hook activity in real time under the condition that the hook static time during the operation is less than the standard value; step 5, monitoring whether the hook moves the casing according to the regulation during the grouting under the condition that the lowering speed during the hook activity is less than the standard value; and step 6, judging whether the casing running process comprehensive quality is satisfied in real time according to the obtained detection information and the actual data. The casing running process comprehensive quality evaluation method can cope with the variable conditions of the operation site through multi-layer judgment, and solves the problem that the traditional method is difficult to accurately determine under complex conditions.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a comprehensive quality assessment method for the cannula insertion process. Background Technology

[0002] Comprehensive quality assessment of the casing running process refers to the real-time monitoring of the casing running stage during various drilling or production processes to identify deviations from normal operating procedures or expected results. This monitoring and identification process aims to quickly detect problems so that timely measures can be taken to prevent potential failures, delays, or safety risks. Casing running data monitoring technology is commonly used in oil and gas exploration, drilling operations, and other fields to ensure the smooth progress of casing running operations and the stability and safety of the overall system. By monitoring real-time data and accurately capturing anomalies during the casing running process, the probability of accidents can be effectively reduced, operational efficiency improved, and production safety guaranteed.

[0003] Data monitoring is particularly important in oil and gas drilling operations because any anomaly can lead to serious safety accidents and economic losses. Existing data monitoring methods mainly include manual detection and threshold alarms. Manual monitoring relies on the operator's experience and judgment, identifying anomalies by observing instrument readings and on-site conditions. This method is highly dependent on the operator's professional knowledge and reaction speed. Preset threshold alarms, on the other hand, involve setting preset thresholds for key parameters (such as pressure, temperature, and flow rate). When the actual parameter exceeds these thresholds, an alarm is triggered. This method is relatively simple, but often only detects obvious anomalies with limited specific conditions.

[0004] While existing data monitoring methods have improved operational safety and efficiency to some extent, several problems remain. Traditional methods rely on post-event analysis or manual judgment, leading to a delay in anomaly identification, which often lags behind the actual occurrence of anomalies and hinders timely warnings and responses. Manual monitoring and simple threshold alarms are inefficient and error-prone in handling complex and changing field environments, failing to fully meet the demands for efficient and safe operations. Furthermore, both methods rely on either manual intervention or static preset values, resulting in insufficient accuracy and potential for missed or false alarms, impacting the reliability of judgments.

[0005] Existing technologies for data monitoring in oil and gas drilling operations suffer from drawbacks such as lag, low efficiency, susceptibility to errors, insufficient accuracy, dependency, limitations, and inability to handle complex and ever-changing anomalies. Traditional data monitoring methods (such as manual detection and threshold alarms) rely on post-event analysis or manual judgment, resulting in anomaly identification often lagging behind the actual occurrence time, failing to provide timely warnings and responses. In oil and gas drilling operations, anomalies are often complex and changeable, and traditional data monitoring methods struggle to comprehensively and accurately capture all types of anomalies, posing significant safety hazards.

[0006] Chinese patent application CN117868689A discloses a casing rotation and circulation running device. The device comprises a data acquisition system (collecting sensor data related to parameters during casing running), a data processing system (processing and analyzing the collected sensor data for subsequent artificial intelligence algorithms), an AI detection system (using AI algorithms to analyze and detect the processed data to monitor abnormalities or problems during casing running in real time), a control system (automatically adjusting and controlling the parameters of the casing rotation and circulation running device based on the AI ​​detection results), and a visualization system (displaying the monitoring, control, and analysis results to the operator in an intuitive manner). This invention, through the AI ​​detection system, can monitor abnormalities during casing running in real time, promptly identify and resolve potential problems, avoid operational risks, and reduce downhole accidents. This application requires a data acquisition system and a control system; the selected model is an LSTM model, which has a relatively complex structure.

[0007] Chinese patent application CN109931048A discloses a method and evaluation system for detecting and evaluating the integrity of oil and gas well casing. The system includes: a wellbore trajectory processing module for horizontal wells in complex conditions; a stress analysis and strength verification module for running completion tubing in horizontal wells in complex conditions; a stress analysis and strength verification module for running operation tubing in horizontal wells in complex conditions; a stress analysis and strength verification module for the fracturing and acidizing process of completion operation tubing in multi-packer horizontal wells; a module for visualizing calculation results and automatically generating reports; and a casing string load analysis and optimization module. The detection method includes: establishing a mechanical model; establishing an evaluation system; implementing operational calculations; casing strength design; establishing a casing string load analysis and optimization module; and data analysis and processing of tubing, casing cement sheath, and formation. The advantages of this invention are its comprehensive, systematic, and scientific evaluation, solving the problem that current methods for analyzing wellbore integrity only consider a single factor. However, this application is only applicable to horizontal wells and cannot be effective for all well types. It can only be used for casing string strength and optimization design, and the results can only be written to a Word document, limiting its application scope.

[0008] Chinese patent application CN114233277B discloses a cementing quality evaluation method based on the attenuation coefficient of casing well mode waves. The method involves: running an acoustic logging instrument into the casing well; transmitting a probe to generate vibration; receiving an array of probes in the well fluid to receive logging waveforms at different source distances; taking the first wave of eight received waveforms; and using a complex exponential modeling method to obtain the time difference dispersion curve and attenuation coefficient dispersion curve of the casing well mode wave. The time difference dispersion curve and attenuation coefficient dispersion curve are projected onto the time difference axis and attenuation coefficient axis, respectively, to obtain the time difference distribution and attenuation coefficient distribution in different frequency ranges. The peak values ​​of each distribution are used to determine the time difference and attenuation coefficient for each frequency range, resulting in time difference variation curves and attenuation coefficient variation curves with depth. Based on the time difference distribution, it is determined whether the first wave is a casing wave. If so, the casing wave amplitude and the corresponding attenuation coefficient are used to evaluate the cementing quality. This invention comprehensively evaluates cementing quality using casing wave amplitude, attenuation coefficient, and time difference distribution. The application evaluates cementing quality, mainly the bonding of cement injected into the annulus after casing is run in, and cannot evaluate the quality of each stage and system of the casing running process.

[0009] The existing technologies described above are significantly different from this invention. A search reveals no literature of the XY category, indicating that this invention is innovative. Since no solution exists in the existing technology to address the technical problem we seek, we have invented a new comprehensive quality assessment method for the cannula insertion process. Summary of the Invention

[0010] The purpose of this invention is to provide a comprehensive quality assessment method for the casing lowering process that can monitor data such as hook status, drill bit position, pump flow, and casing in real time, and issue warnings when abnormalities occur, thereby ensuring the safety of the work site.

[0011] The objective of this invention can be achieved through the following technical measures: a comprehensive quality assessment method for the casing installation process, which includes:

[0012] Step 1: Collect casing data and equipment data for the upcoming operation, as well as real-time data from the work site;

[0013] Step 2: Input the data into the two-way time series model, monitor the process data in real time, and evaluate the data content;

[0014] Step 3: Monitor the time the hook remains stationary during casing operations in real time;

[0015] Step 4: Under the condition that the static time of the hook is less than the standard value during operation, monitor the lowering speed of the hook during movement in real time;

[0016] Step 5: Under the condition that the lowering speed of the hook during its movement is less than the standard value, monitor in real time whether the hook moves the sleeve as required during grouting;

[0017] Step 6: Based on the obtained detection information and actual data, determine in real time whether the overall quality of the insertion process is met.

[0018] The objective of this invention can also be achieved through the following technical measures:

[0019] In step 1, the collected casing and equipment data for the upcoming operation include casing inlet flow rate, casing outer diameter, drill bit size, and maximum drill radius data.

[0020] In step 2, the bidirectional time series model includes:

[0021] The data processing module preprocesses the real-time data from the casing installation site and organizes it into a time series format suitable for input to a bidirectional LSTM network.

[0022] The bidirectional LSTM network module analyzes and models the processed time series data, extracts important temporal feature information, and outputs detection information;

[0023] The judgment and storage module determines whether the overall quality of the loading process is met based on the detection information, and stores the detection and judgment information.

[0024] In step 2, the bidirectional LSTM network module includes:

[0025] The backbone network captures the temporal characteristics and trends in the data;

[0026] The neck network integrates feature information from different time steps of the backbone network to improve the accuracy and robustness of the model;

[0027] The detection network uses the integrated feature information for classification and regression to output detection results for real-time data;

[0028] The training optimization unit adjusts the network parameters based on the training results, selects an appropriate loss function, and chooses the model with the best training results as the final model.

[0029] Step 3 includes:

[0030] Step 301: Process the casing operation-related data collected at the work site and the real-time data to remove noise and invalid data, handle missing values, and ensure the integrity and accuracy of the data; label abnormal situations based on historical data and expert knowledge to provide labels for model training; select appropriate time steps and feature quantities to divide the data into training, validation, and test sets;

[0031] Step 302: Construct a Long Short-Term Memory (LSTM) deep learning time series model in the PyTorch framework, design the network architecture, and adjust the network parameters through optimization algorithms; use cross-validation to optimize hyperparameters such as learning rate, batch size, and number of bidirectional LSTM layers.

[0032] Step 303: Evaluate model performance using accuracy, recall, and F1 score metrics; evaluate the model's generalization ability on the validation and test sets to ensure that the model can effectively identify anomalies.

[0033] Step 304: Apply the trained bidirectional time series model to the real-time data stream to monitor anomalies in the data feed process in real time; based on historical data and the current state, detect whether the current data meets the quality assessment and provide early warning.

[0034] Step 4 includes:

[0035] Step 401: Query the static threshold time based on the mud system and wellbore size used;

[0036] Step 402: By reading real-time data, if the time during which the height of the hook has not changed exceeds the static threshold time period, it is determined that the hook is stationary and an alarm is issued.

[0037] Step 401 specifically includes:

[0038] (41) Obtain the mud system and wellbore size from the casing file, and query the static threshold time;

[0039] (42) Check if there is real-time data input. If there is, proceed to step (43). If not, end the detection.

[0040] (43) Obtain the height difference and time difference of the large hook in the real-time data;

[0041] (44) The time the hook remains stationary is calculated by the height difference and time difference of the hook.

[0042] Step 5 includes:

[0043] Step 501: Calculate the lowering speed of the weak strata based on the relevant data collected at the work site;

[0044] Step 502: If the data item corresponding to the inlet flow is empty in the work site data file, the parameter is obtained by querying based on the drill bit size. If it cannot be found, it is determined that the data item is missing and the inlet flow cannot be obtained.

[0045] Step 503: The maximum allowable speed for casing running is the smaller of the standard value and the running speed in the weak formation. The standard value is determined based on the wellbore size. When data is missing and the inlet flow rate cannot be obtained, the standard value is directly used as the maximum speed limit.

[0046] Step 504: Based on the calculated casing speed limit, monitor the casing installation operation in real time and issue an alarm when the speed exceeds the limit.

[0047] In step 501, the formula for calculating the lowering velocity of the weak strata is:

[0048]

[0049] Among them, Q max D represents the inflow rate, measured in L / s. c D is the outer diameter of the casing, in mm. h D is the wellbore diameter in mm. collar The maximum drill bit radius, in mm, is the largest value obtained from the drill bit assembly excluding the PDC bit; v maxl The maximum permissible rate of lowering the weak strata.

[0050] In step 6, based on the obtained detection information and actual data, it is determined in real time whether the overall quality of the loading process is met, and the detection and judgment information is stored. If it is not met, an alarm is issued.

[0051] Step 6 includes:

[0052] Step 601: Calculate the cumulative length of each casing based on the casing serial number and casing length in the casing record; determine the casing information being lowered based on the field data and casing serial number; look up the elastic modulus based on the casing steel grade and size; obtain the conversion ratio between the cumulative casing length and the casing expansion / contraction based on the elastic modulus; and calculate the casing expansion / contraction.

[0053] Step 602: If the main pump flushing data has data, it is determined that grouting is in progress;

[0054] Step 603: If grouting is in progress, calculate the distance of the hook's single movement based on the real-time data of the hook height and the corresponding time. The value is the maximum distance the hook can continuously move in the same direction without changing direction. When this value is less than twice the sleeve extension / retraction amount, it is determined that an abnormality has occurred and an alarm is issued.

[0055] The objective of this invention can also be achieved through the following technical measures: a comprehensive quality assessment system for the casing insertion process. This system uses a comprehensive quality assessment method for the casing insertion process and sets comprehensive judgment conditions to detect anomalies in the casing insertion process under complex conditions.

[0056] The comprehensive quality assessment method for the casing lowering process in this invention, based on real-time collected work site data, enables real-time monitoring of casing overspeed, hook stationaryness, and abnormalities in the moving casing, and detects whether the current data meets the quality assessment requirements. This invention, by setting comprehensive judgment conditions, enables the system to achieve efficient anomaly detection in complex environments. Compared to the simple judgment relying on a single threshold on traditional equipment, this invention can handle the changing conditions at the work site through multi-layered judgments, solving the problem of traditional methods' difficulty in accurately judging under complex conditions.

[0057] During operation, the system can automatically analyze and identify anomalies, improving detection accuracy. Compared to manual monitoring, the system not only reduces false alarms and missed alarms but also significantly shortens response time, enabling real-time response. After data is uploaded to cloud storage, centralized monitoring is possible, providing management with a real-time and comprehensive monitoring view, which helps improve the overall safety of the work site.

[0058] The system's cloud-based data storage capabilities support subsequent data analysis and model optimization. The accumulation of historical data can be used to build more comprehensive models to further reduce false alarms and false negatives, and even gradually achieve adaptive optimization to cope with the ever-changing needs of the work site. Attached Figure Description

[0059] Figure 1 This is a flowchart of a specific embodiment of the comprehensive quality assessment method for the cannula insertion process of the present invention.

[0060] Figure 2 This is a framework diagram for casing operation data detection in a specific embodiment of the present invention;

[0061] Figure 3 This is a flowchart of the still time detection process in a specific embodiment of the present invention;

[0062] Figure 4 This is a flowchart of the lower sleeve overspeed detection process in a specific embodiment of the present invention;

[0063] Figure 5 This is a flowchart of the movable sleeve detection process in a specific embodiment of the present invention; Detailed Implementation

[0064] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0065] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments of the present invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, and / or combinations thereof.

[0066] like Figure 1 As shown, Figure 1 This is a flowchart of the comprehensive quality assessment method for the casing insertion process according to the present invention. The comprehensive quality assessment method for the casing insertion process includes:

[0067] (1) Collect casing data and equipment data for the upcoming operation in advance, including casing inlet flow rate, casing outer diameter, drill bit size, and maximum drill radius data, and store them in a data file;

[0068] (2) Collect real-time data from the work site.

[0069] (3) Input the data into the two-way time series model, monitor the process data in real time, evaluate the data content, and issue an alarm for data that does not meet the quality assessment.

[0070] (4) Monitor the time the hook remains stationary during casing operation in real time.

[0071] (5) When the time the hook remains stationary is less than the standard value during operation, monitor the lowering speed of the hook during operation in real time.

[0072] (6) Under the condition that the lowering speed of the hook is less than the standard value, monitor in real time whether the hook moves the sleeve as required during grouting.

[0073] (7) Based on the obtained detection information and actual data, determine in real time whether the overall quality of the loading process is met, and store the detection and judgment information. If it is not met, issue an alarm.

[0074] The specific implementation steps of step (3) are as follows:

[0075] (301) Processing: Combine existing on-site casing operation-related data with real-time data, remove noise and invalid data, and handle missing values ​​to ensure data integrity and accuracy. Label anomalies based on historical data and expert knowledge to provide labels for bidirectional time series model training. Select appropriate time steps and feature quantities to divide the data into training, validation, and test sets.

[0076] (302) Construct a long short-term memory network deep learning time series model, i.e., a bidirectional time series model, under the PyTorch framework, design the network architecture, and adjust the network parameters through optimization algorithms. Use methods such as cross-validation to optimize hyperparameters such as learning rate, batch size, and number of bidirectional LSTM layers.

[0077] (303) Use metrics such as accuracy, recall, and F1 score to evaluate the performance of the bidirectional time series model. Evaluate the model's generalization ability on the validation and test sets to ensure that the model can effectively identify anomalies.

[0078] (304) The trained bidirectional time series model is applied to the real-time data stream to monitor anomalies during the data feed. Based on historical data and the current state, the model detects whether the current data meets the quality assessment and provides early warnings.

[0079] The specific implementation steps of step (4) are as follows:

[0080] (401) Query the static threshold time according to the mud system and wellbore size used. The static threshold for running a 139.7mm casing into a 215.9mm wellbore with water-based drilling fluid is 3min.

[0081] (402) By reading real-time data, if the time during which the height of the hook does not change exceeds the static threshold time period, it is determined that the hook is stationary and an alarm is issued.

[0082] The specific implementation steps of step (5) are as follows:

[0083] (501) Calculate the lowering speed of the weak strata based on the relevant data collected at the work site using the following formula:

[0084]

[0085] Among them, Q max D represents the inflow rate, measured in L / s. c D is the outer diameter of the casing, in mm. h D is the wellbore diameter in mm. collar The maximum drill bit radius, in mm, is the largest value obtained from the drill bit assembly excluding the PDC bit; v maxl The maximum permissible rate of lowering into the weakest strata;

[0086] (502) If the data item corresponding to the inlet flow rate is empty in the data file of the work site, the parameter is obtained by querying according to the drill bit size. If it cannot be queried, it is determined that the data item is missing and the inlet flow rate cannot be obtained.

[0087] (503) The maximum allowable speed for running casing is the smaller of the standard value and the speed of running casing in weak formations. The standard value is determined according to the wellbore size. For example, the speed for running 139.7mm casing in a 215.9mm wellbore is 1.2m / s. When the inlet flow rate cannot be obtained due to missing data, the standard value is directly used as the maximum speed limit.

[0088] (504) Based on the calculated casing speed limit, monitor the casing operation in real time and issue an alarm when the speed exceeds the limit.

[0089] To achieve the detection of movable sleeves, the present invention adopts the following solution:

[0090] (601) Calculate the cumulative length of each casing according to the casing serial number and casing length in the casing record; determine the casing information being lowered according to the field data and casing serial number; look up the elastic modulus according to the casing steel grade and size; obtain the conversion ratio between the cumulative length of the casing and the casing expansion and contraction according to the elastic modulus; and calculate the casing expansion and contraction.

[0091] (602) If the “Main Pump Flushing Data” data item has data, it is determined that grouting is in progress;

[0092] (603) If grouting is in progress, the distance of the hook's single movement is calculated based on the hook height and corresponding time in the real-time data. The value is the maximum distance the hook can continuously move in the same direction without changing direction. When this value is less than twice the sleeve extension and retraction, it is determined that an abnormality has occurred and an alarm is issued.

[0093] In summary, the present invention has the following main technical advantages:

[0094] 1. By employing real-time data processing and intelligent discrimination methods, combined with preset discrimination logic or algorithms, real-time monitoring of the casing installation speed, hook stationary time, grouting time, and hook activity during the casing installation process is achieved. This method significantly improves the timeliness and accuracy of discrimination, effectively solving the problems of untimely manual monitoring and insufficient viewing angle.

[0095] 2. Employing real-time alarms and cloud data push, this method can immediately trigger real-time alarms when data anomalies occur during the installation of the bushing, and supports pushing relevant data to the cloud. This real-time alarm and cloud data push mechanism helps to promptly detect and handle problems, improving the efficiency and accuracy of supervision.

[0096] 3. Deep learning models, particularly bidirectional Long Short-Term Memory (LSTM) networks, are introduced to train and analyze historical and real-time data. By learning the temporal characteristics of various parameters during the casing installation process, the bidirectional LSTM model can effectively detect the possibility of anomalies by capturing the correlation between time-series data, thus achieving early warning of abnormal events. The application of this technology makes data analysis in complex environments more accurate and significantly improves the intelligence level of anomaly detection.

[0097] 4. By combining multimodal data fusion technology, data from casing operations, such as drill bit position, hook height, and grouting time, are integrated to form a comprehensive operational status assessment model. Through comprehensive analysis of multiple data sources, the model can more fully reflect the risks and potential problems during the operation, providing a reliable basis for decision-making. This fusion method improves data utilization efficiency and enhances the system's intelligent monitoring capabilities.

[0098] The following are several specific embodiments of the application of the present invention.

[0099] Example 1

[0100] like Figure 2 As shown, the present invention includes a framework diagram for a comprehensive quality time series model detection of the casing insertion process, comprising:

[0101] The data processing module is responsible for preprocessing the real-time data from the casing installation site and organizing it into a time series format suitable for bidirectional LSTM network input.

[0102] The LSTM network module analyzes and models the processed time series data, aiming to extract important time series feature information;

[0103] The judgment and storage module determines whether the overall quality of the insertion process is met based on the detection information, and stores the detection and judgment information.

[0104] In this embodiment, the LSTM network module includes a backbone network, a neck network, a detection network, and a training optimization unit.

[0105] The backbone network is used to capture temporal features and trends in the data. This process helps extract dynamic information that is important for subsequent detection tasks, providing the model with strong contextual understanding capabilities.

[0106] The neck network integrates feature information from different time steps of the backbone network to improve the accuracy and robustness of the model.

[0107] The detection network uses the integrated feature information for classification and regression to output detection results for real-time data.

[0108] The training optimization unit adjusts the network parameters based on the training results, selects an appropriate loss function, and chooses the model with the best training results as the final model.

[0109] In this embodiment, the determination and storage module determines whether the overall quality of the casing installation process is met after the casing operation detection model outputs information, and stores the detection information and determination information.

[0110] Example 2

[0111] like Figure 3 As shown, the process of detecting the static time of the hook height during operation includes the following steps:

[0112] (1) Obtain the mud system and wellbore size from the casing file, and query the static threshold time.

[0113] (2) Check if there is real-time data input. If there is, proceed to step (3). If not, end the detection.

[0114] (3) Obtain the height difference and time difference of the large hook in the real-time data.

[0115] (4) The time the hook remains stationary is calculated by using the height difference and time difference of the hook.

[0116] (5) If the height of the hook remains unchanged for longer than the static threshold time, an alarm will be triggered; return to step (2).

[0117] Example 3

[0118] like Figure 4 As shown, the casing speed monitoring process during operation includes the following steps:

[0119] (1) Calculate the wellbore diameter D by multiplying the drill bit size obtained from the casing record by 105%. h Obtain the outer diameter D of the casing from the casing record. c Obtain the maximum value of the inflow traffic from the data table, convert it to L / s, and denote it as Q. max The largest value obtained from the drill string assembly, excluding the PDC drill bit, is denoted as the maximum drill string radius D. Collor Calculate the lowering rate in the weak formation; based on the wellbore diameter D h Query the standard value for the speed of data transfer.

[0120] (2) Take the minimum value between the lowering speed of the weak stratum and the standard value as the maximum lowering speed.

[0121] (3) Check if there is real-time data input. If there is, proceed to step (4). If not, end the detection.

[0122] (4) Obtain the height difference and time difference of the hook from the real-time data, and calculate the speed of the casing.

[0123] (5) Compare the lowering speed of the casing with the maximum lowering speed to determine if it exceeds the speed limit. If it exceeds the speed limit, issue an alarm; return to step (3).

[0124] Example 4

[0125] like Figure 5 As shown, the process of inspecting the movable casing during grouting includes the following steps:

[0126] (1) Calculate the cumulative length of each casing according to the casing serial number and casing length in the casing record; look up the elastic modulus according to the casing steel grade and size, obtain the conversion ratio between the cumulative length of the casing and the casing expansion and contraction according to the elastic modulus, and calculate the casing expansion and contraction.

[0127] (2) Check if there is real-time data input. If there is, proceed to step (3). If not, end the detection.

[0128] (3) Check whether the real-time data contains the total pump flush. If it does, proceed to step (4); otherwise, return to step (2).

[0129] (4) Calculate the one-time movement distance of the hook based on the hook height in the real-time data to obtain the lifting and lowering distance of the hook;

[0130] (5) Determine whether the distance of the hook being lifted or lowered once is greater than twice the current sleeve extension / retraction amount. If it is less than twice, issue an alarm; return to step (2).

[0131] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0132] Except for the technical features described in the specification, all other technologies are known to those skilled in the art.

Claims

1. A comprehensive quality assessment method for the casing installation process, characterized in that, The comprehensive quality assessment method for the casing installation process includes: Step 1: Collect casing data and equipment data for the upcoming operation, as well as real-time data from the work site; Step 2: Input the data into the two-way time series model, monitor the process data in real time, and evaluate the data content; Step 3: Monitor the time the hook remains stationary during casing operations in real time; Step 4: Under the condition that the static time of the hook is less than the standard value during operation, monitor the lowering speed of the hook during movement in real time; Step 5: Under the condition that the lowering speed of the hook during its movement is less than the standard value, monitor in real time whether the hook moves the sleeve as required during grouting; Step 6: Based on the obtained detection information and actual data, determine in real time whether the overall quality of the insertion process is met.

2. The comprehensive quality assessment method for the casing insertion process according to claim 1, characterized in that, In step 1, the collected casing and equipment data for the upcoming operation include casing inlet flow rate, casing outer diameter, drill bit size, and maximum drill radius data.

3. The comprehensive quality assessment method for the casing insertion process according to claim 1, characterized in that, In step 2, the bidirectional time series model includes: The data processing module preprocesses the real-time data from the casing installation site and organizes it into a time series format suitable for input to a bidirectional LSTM network. The bidirectional LSTM network module analyzes and models the processed time series data, extracts important temporal feature information, and outputs detection information; The judgment and storage module determines whether the overall quality of the loading process is met based on the detection information, and stores the detection and judgment information.

4. The comprehensive quality assessment method for the casing insertion process according to claim 3, characterized in that, In step 2, the bidirectional LSTM network module includes: The backbone network captures the temporal characteristics and trends in the data; The neck network integrates feature information from different time steps of the backbone network to improve the accuracy and robustness of the model; The detection network uses the integrated feature information for classification and regression to output detection results for real-time data; The training optimization unit adjusts the network parameters based on the training results, selects an appropriate loss function, and chooses the model with the best training results as the final model.

5. The comprehensive quality assessment method for the casing installation process according to claim 1, characterized in that, Step 3 includes: Step 301: Process the casing operation-related data collected at the work site and the real-time data to remove noise and invalid data, handle missing values, and ensure the integrity and accuracy of the data; label abnormal situations based on historical data and expert knowledge to provide labels for model training; select appropriate time steps and feature quantities to divide the data into training, validation, and test sets; Step 302: Construct a Long Short-Term Memory (LSTM) deep learning time series model in the PyTorch framework, design the network architecture, and adjust the network parameters through optimization algorithms; use cross-validation to optimize hyperparameters such as learning rate, batch size, and number of bidirectional LSTM layers. Step 303: Evaluate model performance using accuracy, recall, and F1 score metrics; evaluate the model's generalization ability on the validation and test sets to ensure that the model can effectively identify anomalies. Step 304: Apply the trained bidirectional time series model to the real-time data stream to monitor anomalies in the data feed process in real time; based on historical data and the current state, detect whether the current data meets the quality assessment and provide early warning.

6. The comprehensive quality assessment method for the casing installation process according to claim 1, characterized in that, Step 4 includes: Step 401: Query the static threshold time based on the mud system and wellbore size used; Step 402: By reading real-time data, if the time during which the height of the hook has not changed exceeds the static threshold time period, it is determined that the hook is stationary and an alarm is issued.

7. The comprehensive quality assessment method for the casing installation process according to claim 6, characterized in that, Step 401 specifically includes: (41) Obtain the mud system and wellbore size from the casing file, and query the static threshold time; (42) Check if there is real-time data input. If there is, proceed to step (43). If not, end the detection. (43) Obtain the height difference and time difference of the large hook in the real-time data; (44) The time the hook remains stationary is calculated by the height difference and time difference of the hook.

8. The comprehensive quality assessment method for the casing installation process according to claim 1, characterized in that, Step 5 includes: Step 501: Calculate the lowering speed of the weak strata based on the relevant data collected at the work site; Step 502: If the data item corresponding to the inlet flow is empty in the work site data file, the parameter is obtained by querying based on the drill bit size. If it cannot be found, it is determined that the data item is missing and the inlet flow cannot be obtained. Step 503: The maximum allowable speed for casing running is the smaller of the standard value and the running speed in the weak formation. The standard value is determined based on the wellbore size. When data is missing and the inlet flow rate cannot be obtained, the standard value is directly used as the maximum speed limit. Step 504: Based on the calculated casing speed limit, monitor the casing installation operation in real time and issue an alarm when the speed exceeds the limit.

9. The comprehensive quality assessment method for the casing installation process according to claim 8, characterized in that, In step 501, the formula for calculating the lowering velocity of the weak strata is: Among them, Q max D represents the inflow rate, measured in L / s. c D is the outer diameter of the casing, in mm. h D is the wellbore diameter in mm. collar The maximum drill bit radius, in mm, is the largest value obtained from the drill bit assembly excluding the PDC bit; v maxl The maximum permissible rate of lowering the weak strata.

10. The comprehensive quality assessment method for the casing installation process according to claim 1, characterized in that, In step 6, based on the obtained detection information and actual data, it is determined in real time whether the overall quality of the loading process is met, and the detection and judgment information is stored. If it is not met, an alarm is issued.

11. The comprehensive quality assessment method for the casing insertion process according to claim 10, characterized in that, Step 6 includes: Step 601: Calculate the cumulative length of each casing based on the casing serial number and casing length in the casing record; determine the casing information being lowered based on the field data and casing serial number; look up the elastic modulus based on the casing steel grade and size; obtain the conversion ratio between the cumulative casing length and the casing expansion / contraction based on the elastic modulus; and calculate the casing expansion / contraction. Step 602: If the main pump flushing data has data, it is determined that grouting is in progress; Step 603: If grouting is in progress, calculate the distance of the hook's single movement based on the real-time data of the hook height and the corresponding time. The value is the maximum distance the hook can continuously move in the same direction without changing direction. When this value is less than twice the sleeve extension / retraction amount, it is determined that an abnormality has occurred and an alarm is issued.

12. A comprehensive quality assessment system for the casing installation process, characterized in that, The comprehensive quality assessment system for the casing insertion process adopts the comprehensive quality assessment method for the casing insertion process described in any one of claims 1-11. By setting comprehensive judgment conditions, it realizes the abnormal detection of the casing insertion process in complex environments.

Images