Early kick monitoring method and system for drilling

Abstract

The embodiments of the present invention belong to the technical field of kick monitoring. Provided are an early kick monitoring method and system for drilling. The method comprises: collecting, in real time, a drilling fluid recovery state parameter in a drilling process; using the preprocessed drilling fluid recovery state parameter to invoke a pre-trained kick determination model, and executing the kick determination model to determine whether there is a kick, wherein the kick determination model is obtained by means of performing training on the basis of stored drilling data on which kick development state calibration has been executed; when it is determined that there is a kick, calculating flow-rate fluctuation rates on the basis of drilling fluid flow rates in a kick state within preset sampling periods, generating kick trends on the basis of the flow-rate fluctuation rates, and comparing the kick trends within the preset sampling periods to obtain a kick development state; and on the basis of the kick development state generated in real time, matching and executing a corresponding response scheme. The solution in the present invention solves the problems of low accuracy and low response speed in the existing early kick monitoring solutions.

Description

Early-stage wellbore overflow monitoring methods and systems

[0001] Cross-reference to related applications

[0002] This application claims the benefit of Chinese Patent Application No. 202411210574.1, filed on August 30, 2024, the contents of which are incorporated herein by reference. Technical Field

[0003] This invention relates to the field of overflow monitoring technology, specifically to a method and system for monitoring early-stage well overflows. Background Technology

[0004] In oil exploration and development, when drilling penetrates an oil and gas reservoir, fluids from the formation can infiltrate the drilling fluid and return to the surface along with it; this process is called a blowout. If a blowout is not detected and dealt with in time, it can lead to a well kick, a blowout, or even an uncontrolled blowout, causing a serious catastrophic accident. To avoid such accidents, the most effective method is to detect blowouts promptly and handle them effectively using technical means. However, there are still some shortcomings in early blowout monitoring and prediction, mainly in the following five aspects:

[0005] 1) Current common overflow early warning methods often only confirm the overflow after the overflow phenomenon has become quite obvious. This lag means that the overflow downhole has already developed to a certain extent, increasing the difficulty and risk of handling it. An ideal early warning system should be able to detect anomalies in the early stages of an overflow, thus gaining valuable time to take preventive measures.

[0006] 2) Currently, most methods for determining a blowout rely primarily on the analysis of parameters such as drilling fluid volume and level. While changes in these parameters can reflect the presence of a blowout, they are too simplistic and lack comprehensiveness. Changes in drilling fluid volume and level can be influenced by a variety of factors, making it difficult to accurately determine a blowout based solely on these parameters, and misjudgments are prone to occur.

[0007] 3) Existing overflow monitoring methods typically require manual inspection to determine whether an overflow has occurred. This not only increases the complexity and workload of the operation but also increases the risk of misjudgment due to human factors. The manual inspection process is also time-consuming and fails to fully utilize automation and intelligent technologies for real-time monitoring and analysis.

[0008] 4) Existing overflow early warning systems have a certain time lag, failing to provide sufficient time for the implementation of control measures. Once an overflow occurs, control time is critical, and the lag in the early warning system may lead to missed opportunities for optimal control, increasing the risk of accidents. Early warning systems should be real-time and efficient to ensure rapid response and handling.

[0009] 5) Existing methods struggle to effectively eliminate false overflows caused by manual operation or reporting during the overflow identification process, resulting in a high false alarm rate. These false alarms not only waste human and material resources but also disrupt normal construction progress. Frequent false alarms reduce staff's trust in the early warning system, thus affecting its effectiveness in actual operation.

[0010] In summary, the aforementioned shortcomings in early overflow monitoring and prediction all contribute to increased well control safety risks. Therefore, a new early overflow monitoring method is needed to address this issue. Summary of the Invention

[0011] The purpose of this invention is to provide a method and system for monitoring early-stage well overflows, so as to at least solve the problems of low accuracy and slow response speed of existing early-stage well overflow monitoring schemes.

[0012] To achieve the above objectives, the first aspect of the present invention provides a method for monitoring early-stage wellbore overflow, the method comprising: real-time acquisition of drilling fluid recovery status parameters during the drilling process based on a backflow sampling device, and preprocessing the drilling fluid recovery status parameters; calling a pre-trained overflow judgment model on the preprocessed drilling fluid recovery status parameters, and executing the overflow judgment model to determine whether an overflow exists; wherein the overflow judgment model is obtained by training on existing drilling data after overflow development state calibration; when an overflow is determined to exist, calculating the flow rate fluctuation rate based on the drilling fluid flow rate under the overflow state within a preset sampling period, generating an overflow trend based on the flow rate fluctuation rate, and comparing the overflow trends of each preset sampling period to obtain the overflow development state; matching and executing a corresponding response plan based on the real-time generated overflow development state.

[0013] Optionally, the reflux sampling device includes: a device body, including a flow measurement chamber, and multiple flow sensors disposed on the device body; the flow measurement chamber includes an inlet chamber and an outlet chamber; the device body introduces the recovered drilling fluid through a drilling fluid inlet; a quantitative sampler and flow sensor A are disposed on the pipeline between the drilling fluid inlet and the inlet chamber of the flow measurement chamber; other flow sensors are disposed at different height positions within the flow measurement chamber to monitor the drilling fluid flow rate under different flow rate conditions.

[0014] Optionally, the drilling fluid recovery status parameters include: recovered drilling fluid flow rate, inlet fluid level, outlet fluid level, recovered drilling fluid density, recovered drilling fluid temperature, recovered drilling fluid conductivity, and recovered drilling fluid gas content.

[0015] Optionally, the method further includes: performing pre-training of the overflow judgment model, including: collecting existing drilling data and performing overflow status calibration on each group of drilling data; constructing a first training sample based on the existing drilling data with completed overflow status calibration; training the first training sample in the pre-constructed neural network to obtain an initial overflow judgment model; and performing verification on the initial overflow judgment model based on the first verification sample reserved in the first training sample, and using the model that passes the verification as the overflow judgment model.

[0016] Optionally, the preprocessing of the drilling fluid recovery state parameters includes: sequentially performing filtering, damping, and short-cycle averaging on the drilling fluid recovery state parameters.

[0017] Optionally, the step of using the pre-processed drilling fluid recovery state parameters as input parameters of a pre-trained overflow judgment model, performing model training, and judging whether an overflow exists based on the training results of the overflow judgment model includes: obtaining real-time training results based on real-time drilling fluid recovery state parameters within a preset detection period; judging whether an overflow exists based on the real-time training results, and judging the overflow type under the overflow state.

[0018] Optionally, the step of determining whether an overflow exists based on real-time training results, and determining the overflow type under the overflow state, includes: if the recovered drilling fluid density shows a decreasing trend, the recovered drilling fluid temperature and conductivity show an increasing trend, and the recovered drilling fluid gas content does not show a changing trend, an overflow is determined to exist, and the overflow type is water intrusion; if the recovered drilling fluid density and conductivity show a decreasing trend, and the recovered drilling fluid temperature and gas content do not show a changing trend, an overflow is determined to exist, and the overflow type is oil intrusion; if the recovered drilling fluid density, conductivity, and temperature all show a decreasing trend, and the drilling fluid gas content shows an increasing trend, an overflow is determined to exist, and the overflow type is gas intrusion; other cases are determined to be non-overflowing.

[0019] Optionally, the method further includes: performing pre-training of the overflow development model, including: collecting existing drilling data and calibrating the overflow development state of each group of drilling data; in the drilling data with overflow development state calibration, using drilling fluid flow rate as a key sample and other data besides drilling fluid flow rate as auxiliary samples; constructing a second training sample based on the key sample and the auxiliary sample; training the second training sample in the pre-constructed neural network to obtain an initial overflow judgment model; and verifying the initial overflow judgment model based on the verification samples reserved in the second training sample, and using the model that passes the verification as the overflow judgment model.

[0020] Optionally, when overflow is determined to exist, the step of calculating the flow rate fluctuation rate based on the drilling fluid flow rate under overflow conditions within a preset sampling period, generating an overflow trend based on the flow rate fluctuation rate, and comparing the overflow trends of each preset sampling period to obtain the overflow development state includes: performing filtering and averaging processing on the drilling fluid flow rate under overflow conditions within the preset sampling period to obtain a flow rate filter value and a flow rate average value, respectively; calculating the flow rate fluctuation rate based on the flow rate filter value and the flow rate average value; calculating the overflow trend based on the flow rate fluctuation rate, and comparing the overflow trends of each preset sampling period to obtain the overflow development state.

[0021] Optionally, the calculation rule for the overflow trend is as follows:

[0022] Where EQs represents the overflow trend; i represents the sampling time point; p represents the total sampling points; Qs′=(Q2-Q) / Q*100%, and Qs=|Qs′|, Qs represents the flow fluctuation rate; Q2 represents the flow filter value; and Q represents the average flow rate.

[0023] Optionally, the step of calculating the overflow trend based on the flow fluctuation rate and comparing the overflow trend of each preset sampling period to obtain the overflow development state includes: if the overflow trend of the current sampling period is greater than the overflow trend of the previous sampling period, it indicates that the current overflow development state is rising; if the overflow trend of the current sampling period is less than the overflow trend of the previous sampling period, it indicates that the current overflow development state is falling.

[0024] Optionally, based on the real-time generated overflow development status, a corresponding response plan is matched and executed, including: if the overflow development status is rising, and the duration of the rising status is greater than a preset monitoring time, or the cumulative overflow amount is greater than a preset cumulative overflow amount, then the corresponding response plan is an alarm plan; if the duration of the rising status does not reach the preset monitoring time, or the cumulative overflow amount does not reach the preset cumulative overflow amount, and it is determined that there is no overflow at the current moment, then the corresponding response plan is a zeroing and re-accumulation plan; wherein, the zeroing and re-accumulation plan involves recalculating the cumulative overflow amount and the duration of the rising status; if the overflow development status is falling, then the corresponding response plan is a continuous monitoring plan.

[0025] A second aspect of the present invention provides an early-stage wellbore overflow monitoring system, the system comprising: a data acquisition unit, configured to acquire drilling fluid recovery status parameters in real time during the drilling process based on a backflow sampling device, and to preprocess the drilling fluid recovery status parameters; a judgment unit, configured to call a pre-trained overflow judgment model based on the preprocessed drilling fluid recovery status parameters, and execute the overflow judgment model to determine whether an overflow exists; wherein the overflow judgment model is trained based on existing drilling data after overflow development state calibration; a monitoring unit, configured to calculate the flow fluctuation rate based on the drilling fluid flow rate under the overflow state within a preset sampling period when an overflow is determined to exist, and to generate an overflow trend based on the flow fluctuation rate, and to obtain the overflow development state by comparing the overflow trends of each preset sampling period; and a response unit, configured to match and execute a corresponding response plan based on the real-time generated overflow development state.

[0026] Optionally, the reflux sampling device includes: a device body, including a flow measurement chamber, and multiple flow sensors disposed on the device body; the flow measurement chamber includes an inlet chamber and an outlet chamber; the device body introduces the recovered drilling fluid through a drilling fluid inlet; a quantitative sampler and flow sensor A are disposed on the pipeline between the drilling fluid inlet and the inlet chamber of the flow measurement chamber; other flow sensors are disposed at different height positions within the flow measurement chamber to monitor the drilling fluid flow rate under different flow rate conditions.

[0027] On the other hand, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for monitoring early well overflows.

[0028] The above technical solution involves real-time acquisition of drilling fluid recovery status parameters during the drilling process, followed by preprocessing to ensure data accuracy and reliability. The preprocessed data is then input into a pre-trained overflow detection model for training, enabling accurate identification of overflow phenomena in their initial stages and avoiding accident risks due to delayed warnings. Furthermore, upon detecting an overflow, the system monitors the drilling fluid flow rate in real time and uses it as input to the overflow development model for real-time training and status generation. This approach ensures accurate tracking and prediction of every step in the overflow development process, providing timely and effective response strategies. This technical solution not only improves the accuracy and real-time performance of overflow detection and reduces false alarm rates, but also significantly reduces reliance on manual inspections through automated monitoring and intelligent model analysis, saving manpower and improving operational efficiency.

[0029] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0030] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0031] Figure 1 is a flowchart of the steps of a drilling early overflow monitoring method provided by an embodiment of the present invention;

[0032] Figure 2 is a schematic diagram of the reflux sampling device provided in one embodiment of the present invention;

[0033] Figure 3 is a system structure diagram of an early-stage well overflow monitoring system provided in one embodiment of the present invention.

[0034] Explanation of reference numerals in the attached drawings: 1-Main body of the device; 2-Flow sensor B; 3-Flow sensor A; 4-Flow measurement chamber; 5-Drilling fluid inlet; 6-Quantitative sampler; 7-Flow sensor C; 8-Flow sensor D. Detailed Implementation

[0035] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0036] Figure 1 is a flowchart of a drilling early-stage overflow monitoring method provided by an embodiment of the present invention. As shown in Figure 1, this embodiment of the present invention provides a drilling early-stage overflow monitoring method, the method comprising:

[0037] Step S10: Real-time acquisition of drilling fluid recovery status parameters during the drilling process based on the reflux sampling device, and preprocessing of the drilling fluid recovery status parameters.

[0038] Specifically, the drilling fluid recovery status parameters include: recovered drilling fluid flow rate, inlet fluid level, outlet fluid level, recovered drilling fluid density, recovered drilling fluid temperature, recovered drilling fluid conductivity, and recovered drilling fluid gas content.

[0039] In one possible implementation, as shown in Figure 2, the reflux sampling device includes: a device body 1, including a flow measurement chamber 4, and multiple flow sensors disposed on the device body 1; the flow measurement chamber 4 includes an inlet chamber and an outlet chamber; the device body 1 introduces the recovered drilling fluid through a drilling fluid inlet 5; a quantitative sampler 6 and a flow sensor A3 are disposed on the pipeline between the drilling fluid inlet 5 and the inlet chamber of the flow measurement chamber 4; flow sensors B2, C7, and D8 are disposed at different positions within the flow measurement chamber 4 to monitor the drilling fluid flow rate at different positions. Figure 2 illustrates the use of four sensors. Except for flow sensor A, the other flow sensors are used to monitor the drilling fluid flow rate under different flow rate conditions. Other numbers of sensors can be used as long as the installation heights remain different.

[0040] In this invention, overflow monitoring is a crucial aspect of oil exploration and development. In actual operations, if overflows are not detected and addressed promptly, they can lead to severe well kicks and blowouts, causing significant economic losses and environmental damage. Therefore, accurate and timely overflow monitoring is paramount. Traditional overflow monitoring methods often rely on monitoring a single parameter (such as flow rate Q). While simple, this method has limitations in practical applications, such as the inability to effectively identify early overflows, susceptibility to noise interference, and a high false alarm rate. This invention proposes an overflow monitoring method based on a backflow sampling device and multi-sensor monitoring technology. By collecting drilling fluid recovery status parameters in real time and preprocessing these parameters, the accuracy and reliability of the data are improved. In practical applications, the focus of overflow monitoring is on changes in flow rate Q. However, to improve the accuracy of overflow monitoring, this invention not only monitors flow rate Q but also combines other parameters (such as liquid level, density, temperature, conductivity, and gas content) for comprehensive analysis to determine the actual situation and provide early warnings.

[0041] In one possible implementation, this invention proposes a multi-flow sensor monitoring solution. By installing multiple flow sensors and performing mutual consistency verification, the random errors of a single sensor are avoided, thereby ensuring the accuracy of flow data acquisition. This multi-sensor monitoring technology has the following advantages:

[0042] 1) When multiple sensors collect flow data simultaneously, consistency can be verified by comparing the data from each sensor, identifying and eliminating abnormal data. This method can significantly improve data accuracy and reduce misjudgments caused by the failure or error of a single sensor.

[0043] 2) The system can dynamically adjust the preprocessing algorithm based on real-time monitored sensor data to adapt to different operating environments and conditions. This not only improves the system's flexibility and adaptability but also enhances its ability to identify overflows in complex environments.

[0044] 3) In addition to changes in flow rate Q, the system also monitors other key parameters such as liquid level, density, temperature, conductivity, and gas content. Comprehensive analysis of these parameters can provide a more complete basis for overflow judgment. For example, in the initial stage of overflow, although the change in flow rate Q may not be obvious, parameters such as density, temperature, conductivity, and gas content may undergo slight changes. These changes can serve as important evidence for early overflow judgment.

[0045] Preferably, the preprocessing of the drilling fluid recovery state parameters includes: sequentially performing filtering, damping, and short-cycle averaging on the drilling fluid recovery state parameters.

[0046] In this embodiment of the invention, to further improve data reliability, the collected drilling fluid recovery status parameters undergo preprocessing, including filtering, damping, and short-cycle average calculation. A moving average filter or low-pass filter is used to remove high-frequency noise from the data, making the signal smoother. Exponential smoothing or damping factors are used to reduce the impact of instantaneous spikes or fluctuations, ensuring signal smoothness and continuity. Rolling averaging or sliding window techniques are used to calculate the average data value within a certain time window, further smoothing data changes and removing abnormal fluctuations.

[0047] Step S20: The pre-processed drilling fluid recovery status parameters are called into the pre-trained overflow judgment model, and the overflow judgment model is executed to determine whether an overflow exists.

[0048] Specifically, during normal drilling operations without any abnormalities, the following characteristics exist due to the structural design and parameter characteristics of the non-full pipe measurement device at the drilling fluid outlet:

[0049] 1) Flow rate Q, inlet liquid level L in and outlet liquid level L out Due to various factors such as installation, drilling rig status, blockage, flow channels, and slurry return, there may occasionally be drastic fluctuations and data spikes. Therefore, data preprocessing is required to reduce data spikes (sudden highs and lows) and retain data change characteristics. Instantaneous values ​​are filtered and damped, while the average value over a short period is also used. The specific methods are described later.

[0050] 2) Parameters such as density ρ, temperature T, electrical conductivity E, and gas content G are less affected by factors such as installation, drilling rig status, blockage, flow channels, and slurry return, therefore their fluctuation range is relatively small, fluctuating within a certain range. The relevant characteristics and ranges are obtained through daily operational data analysis and learning.

[0051] Based on this, the method further includes: performing pre-training of the overflow judgment model, including: collecting existing drilling data and performing overflow status calibration on each group of drilling data; constructing a first training sample based on the existing drilling data with completed overflow status calibration; performing training on the first training sample in the pre-constructed neural network to obtain an initial overflow judgment model; performing verification on the initial overflow judgment model based on the first verification sample reserved in the first training sample, and using the model that passes the verification as the overflow judgment model.

[0052] In this embodiment of the invention, existing drilling data is collected, including historical drilling data and simulation experimental data. Historical drilling data comes from records of actual drilling processes and contains a wealth of real-world operational information and overflow cases. Simulation experimental data, on the other hand, is experimental data on overflow phenomena obtained under laboratory conditions by simulating real drilling environments and operations. The combination of both provides rich and comprehensive training samples. Overflow state calibration is performed on each set of collected drilling data. Overflow state calibration refers to determining whether an overflow occurred at each time point in the dataset, and the specific type and extent of the overflow, based on parameter changes during the drilling process. This step is crucial because only data with accurately calibrated overflow states can serve as effective training samples for model construction.

[0053] Furthermore, based on existing drilling data with completed overflow condition calibration, the first training sample is constructed. The training sample includes various parameters, such as flow rate, liquid level, density, temperature, conductivity, and gas content, which exhibit different trends when an overflow occurs. By analyzing the patterns of these parameter changes, sufficient training data can be provided to the neural network, helping the model learn how to identify overflows.

[0054] Furthermore, the first training sample is trained on the pre-built neural network. Neural networks are powerful machine learning algorithms capable of automatically learning complex nonlinear relationships through training on large amounts of data. During training, the neural network adjusts its internal weights, gradually optimizing the model's predictive ability to accurately identify overflow states. After obtaining the initial overflow judgment model, the model is validated using the first validation samples reserved in the training samples. Validation samples are a subset of data separated from the training data, not involved in the training process, and are used to evaluate the model's generalization ability. Validation with validation samples ensures that the model performs well on unseen data, avoiding overfitting. The validated model is then used as the final overflow judgment model. After validation, the model can accurately distinguish between overflow and non-overflow states and identify different types of overflows, providing reliable early warning and guidance for actual drilling operations.

[0055] Furthermore, the step of using the pre-processed drilling fluid recovery state parameters as input parameters for a pre-trained overflow judgment model, performing model training, and judging whether an overflow exists based on the training results of the overflow judgment model includes: obtaining real-time training results based on real-time drilling fluid recovery state parameters within a preset detection period; judging whether an overflow exists based on the real-time training results, and judging the overflow type under the overflow state.

[0056] Specifically, if the recovered drilling fluid density shows a decreasing trend, the recovered drilling fluid temperature and conductivity show an increasing trend, and the recovered drilling fluid gas content shows no changing trend, an overflow is determined to exist, and the overflow type is water intrusion; if the recovered drilling fluid density and conductivity show a decreasing trend, and the recovered drilling fluid temperature and gas content show no changing trend, an overflow is determined to exist, and the overflow type is oil intrusion; if the recovered drilling fluid density, conductivity, and temperature all show a decreasing trend, and the drilling fluid gas content shows an increasing trend, an overflow is determined to exist, and the overflow type is gas intrusion; all other cases are determined not to be an overflow.

[0057] Step S30: When overflow is determined, calculate the flow rate fluctuation rate based on the drilling fluid flow rate under the overflow state within the preset sampling period, generate the overflow trend based on the flow rate fluctuation rate, and obtain the overflow development state by comparing the overflow trend of each preset sampling period.

[0058] Specifically, filtering and averaging are performed on the drilling fluid flow rate under overflow conditions within a preset sampling period to obtain the flow rate filter value and the flow rate average value, respectively; the flow rate fluctuation rate is calculated based on the flow rate filter value and the flow rate average value; the overflow trend is calculated based on the flow rate fluctuation rate, and the overflow development status is obtained by comparing the overflow trend of each preset sampling period.

[0059] Furthermore, the calculation rule for the overflow trend is as follows:

[0060] Where EQs represents the overflow trend; i represents the sampling time point; p represents the total sampling points; Qs′=(Q2-Q) / Q*100%, and Qs=|Qs′|, Qs represents the flow fluctuation rate; Q2 represents the flow filter value; and Q represents the average flow rate.

[0061] Preferably, if the overflow trend of the current sampling period is greater than the overflow trend of the previous sampling period, it indicates that the current overflow development state is rising; if the overflow trend of the current sampling period is less than the overflow trend of the previous sampling period, it indicates that the current overflow development state is falling.

[0062] Step S40: Match and execute the corresponding response plan based on the real-time generated overflow development status.

[0063] Specifically, if the overflow development state is in an upward state, and the duration of the upward state is greater than the preset monitoring time, or the cumulative overflow amount is greater than the preset cumulative overflow amount, the corresponding response plan is an alarm plan; if the duration of the upward state is less than the preset monitoring time, or the cumulative overflow amount is less than the preset cumulative overflow amount, and it is determined that there is no overflow at the current moment, the corresponding response plan is a zeroing and re-accumulation plan; wherein, the zeroing and re-accumulation plan involves recalculating the cumulative overflow amount and the duration of the upward state; if the overflow development state is in a downward state, the corresponding response plan is a continuous monitoring plan.

[0064] In one possible implementation, based on the parameter changes caused by different intruders, when the set warning is exceeded, the current state is recorded and the overflow accumulation amount EYQ is calculated, and the timer M3△t is started. When the state disappears and then recovers, the accumulation and timing need to be restarted.

[0065] Furthermore, as the early overflow continues to develop, the flow rate will increase to varying degrees as the overflow increases (exceeding the normal fluctuation level Qs1; the average flow rate Q and the flow rate ratio QK both show an upward trend and exceed normal levels), while the inlet liquid level L... in and outlet liquid level L out There will also be a significant increase, and all the above parameters have warning values ​​for the range of change. When the warning mechanism is triggered, or M3△t reaches the time limit or the overflow accumulation EYQ reaches the warning value, the status is recorded and an early overflow prediction warning (including intrusion classification information) will be popped up immediately.

[0066] Furthermore, when an overflow occurs, the flow rate increases to varying degrees depending on the magnitude of the overflow YQ (exceeding the normal fluctuation level Qs1; the average flow rate Q and the flow rate ratio QK both show an upward trend and exceed normal levels). The overflow behavior is affected by (oil, gas, and water) intrusions and the overflow volume, and is also influenced by many factors such as installation, drilling rig status, blockage, flow channels, and slurry return. However, the duration Δt and the upward trend of the flow rate can be separated through frequency domain analysis. Simultaneously with the increase in flow rate, the inlet liquid level L... in and outlet liquid level L out An upward trend will appear successively, and at the same time, the slurry density ρ, temperature T, electrical conductivity E, and gas content G will all change to a certain level. Start recording the current state M11 and calculating the overflow accumulation EYQ, and time M11△t. When the state disappears and then recovers, the accumulation and timing need to be restarted.

[0067] Figure 3 is a system structure diagram of an early-stage wellbore overflow monitoring system provided by an embodiment of the present invention. As shown in Figure 3, the embodiment of the present invention provides an early-stage wellbore overflow monitoring system, the system comprising: a data acquisition unit, used to acquire drilling fluid recovery status parameters in real time during the drilling process based on a backflow sampling device, and to perform preprocessing on the drilling fluid recovery status parameters; a judgment unit, used to call a pre-trained overflow judgment model based on the preprocessed drilling fluid recovery status parameters, and execute the overflow judgment model to determine whether an overflow exists; wherein, the overflow judgment model is obtained by training based on existing drilling data after overflow development state calibration; a monitoring unit, used to calculate the flow fluctuation rate based on the drilling fluid flow rate under the overflow state within a preset sampling period when an overflow is determined to exist, and to generate an overflow trend based on the flow fluctuation rate, and to obtain the overflow development state by comparing the overflow trends of each preset sampling period; and a response unit, used to match and execute corresponding response schemes based on the real-time generated overflow development state.

[0068] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for monitoring early well overflows.

[0069] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0070] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0071] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for monitoring early-stage wellbore overflow, characterized in that, The method includes: real-time acquisition of drilling fluid recovery status parameters during the drilling process based on a reflux sampling device, and preprocessing the drilling fluid recovery status parameters; The pre-processed drilling fluid recovery status parameters are used to call a pre-trained overflow detection model, which is then executed to determine whether an overflow exists; wherein, The overflow judgment model is obtained by training existing drilling data after overflow development status calibration. When overflow is detected, the flow rate fluctuation rate is calculated based on the drilling fluid flow rate under the overflow state within the preset sampling period, and the overflow trend is generated based on the flow rate fluctuation rate. The overflow development state is obtained by comparing the overflow trend of each preset sampling period. Match and execute corresponding response plans based on the real-time generated overflow development status.

2. The method according to claim 1, characterized in that, The reflux sampling device includes: The main body of the device includes a flow measurement chamber and multiple flow sensors mounted on the main body of the device; The flow measurement chamber includes an inlet chamber and an outlet chamber; The main body of the device introduces the recovered drilling fluid through the drilling fluid inlet; A quantitative sampler and a flow sensor A are installed on the pipeline between the drilling fluid inlet and the inlet chamber of the flow measurement chamber; Other flow sensors are installed at different locations within the flow measurement chamber to monitor drilling fluid flow at different locations.

3. The method according to claim 2, characterized in that, The drilling fluid recovery status parameters include: The recovered drilling fluid flow rate, inlet fluid level, outlet fluid level, recovered drilling fluid density, recovered drilling fluid temperature, recovered drilling fluid conductivity, and recovered drilling fluid gas content are measured.

4. The method according to claim 1, characterized in that, The method further includes: Perform pre-training of the overflow detection model, including: Collect existing drilling data and perform overflow status calibration on each group of drilling data; The first training sample is constructed based on the existing drilling data that has completed the overflow status calibration; The first training sample is trained in the pre-constructed neural network to obtain an initial overflow judgment model; The initial overflow judgment model is validated based on the first validation sample reserved in the first training sample, and the model that passes the validation is used as the overflow judgment model.

5. The method according to claim 1, characterized in that, The preprocessing of the drilling fluid recovery status parameters includes: The drilling fluid recovery status parameters are sequentially subjected to filtering, damping, and short-cycle averaging.

6. The method according to claim 1, characterized in that, The step of using the pre-processed drilling fluid recovery state parameters as input parameters for a pre-trained overflow detection model, performing model training, and determining whether an overflow exists based on the training results of the overflow detection model includes: Within a preset detection cycle, real-time training results are obtained based on real-time drilling fluid recovery status parameters. The system determines whether an overflow exists based on real-time training results, and then determines the type of overflow in the overflow state.

7. The method according to claim 6, characterized in that, The step of determining whether an overflow exists based on real-time training results, and determining the type of overflow in the overflow state, includes: If the density of the recovered drilling fluid shows a decreasing trend, the temperature and conductivity of the recovered drilling fluid show an increasing trend, and the gas content of the recovered drilling fluid does not show a changing trend, it is determined that there is an overflow, and the overflow type is water intrusion. If the density and conductivity of the recovered drilling fluid show a decreasing trend, while the temperature and gas content of the recovered drilling fluid do not show a changing trend, it is determined that there is an overflow, and the overflow type is oil intrusion. If the density, conductivity, and temperature of the recovered drilling fluid all show a decreasing trend, and the gas content of the drilling fluid shows an increasing trend, it is determined that there is an overflow, and the overflow type is gas intrusion. In other cases, it is determined that there is no overflow.

8. The method according to claim 1, characterized in that, The method further includes: Perform pre-training of the overflow development model, including: Collect existing drilling data and perform overflow development status calibration on each group of drilling data; In drilling data calibrated for overflow development status, drilling fluid flow rate is used as the key sample, and other data besides drilling fluid flow rate are used as auxiliary samples. A second training sample is constructed based on the key sample and the auxiliary sample; The second training sample is trained in the pre-constructed neural network to obtain the initial overflow judgment model; The initial overflow judgment model is validated based on the validation samples reserved in the second training sample, and the model that passes the validation is used as the overflow judgment model.

9. The method according to claim 1, characterized in that, When overflow is determined to exist, the flow rate fluctuation rate is calculated based on the drilling fluid flow rate under overflow conditions within a preset sampling period, and an overflow trend is generated based on the flow rate fluctuation rate. The overflow development state is obtained by comparing the overflow trends of each preset sampling period, including: Filtering and averaging are performed on the drilling fluid flow rate under overflow conditions within a preset sampling period to obtain the filtered flow rate value and the average flow rate value, respectively. The flow fluctuation rate is calculated based on the flow filter value and the average flow rate. The overflow trend is calculated based on the flow fluctuation rate, and the overflow development status is obtained by comparing the overflow trend of each preset sampling period.

10. The method according to claim 9, characterized in that, The calculation rule for the overflow trend is as follows: Among them, EQs represents the overflow trend; Qs′=(Q2-Q) / Q*100%, and Qs=|Qs′|, where Qs is the flow fluctuation rate; i represents the sampling time point; p represents the total number of sampling points; Q2 is the flow filtering value; Q represents the average flow rate.

11. The method according to claim 10, characterized in that, The step of calculating the overflow trend based on the flow fluctuation rate and comparing the overflow trend in each preset sampling period to obtain the overflow development status includes: If the overflow trend in the current sampling period is greater than the overflow trend in the previous sampling period, it indicates that the current overflow development state is on the rise. If the overflow trend in the current sampling period is less than the overflow trend in the previous sampling period, it indicates that the current overflow development state is in a decreasing state.

12. The method according to claim 11, characterized in that, Based on the real-time generated overflow development status, match and execute corresponding response plans, including: If the overflow development status is rising, the duration of the rising status is longer than the preset monitoring time, or the cumulative overflow amount is greater than the preset cumulative overflow amount, then the corresponding response plan is the alarm plan. If the duration of the rising state does not reach the preset monitoring time, or the accumulated overflow does not reach the preset accumulated overflow amount, and it is determined that there is no overflow at the current moment, the corresponding response is a zeroing and re-accumulation scheme; among which, The zeroing and re-accumulation scheme involves recalculating the overflow accumulation and the duration of the rising state. If the overflow is in a declining state, the corresponding response plan is a continuous monitoring plan.

13. A wellbore early overflow monitoring system, characterized in that, The system includes: The acquisition unit is used to acquire drilling fluid recovery status parameters in real time during the drilling process based on the reflux sampling device, and to perform preprocessing on the drilling fluid recovery status parameters. The judgment unit is used to call a pre-trained overflow judgment model based on the pre-processed drilling fluid recovery state parameters, and execute the overflow judgment model to determine whether an overflow exists; wherein, The overflow judgment model is obtained by training existing drilling data after overflow development status calibration. The monitoring unit is used to calculate the flow fluctuation rate based on the drilling fluid flow rate under the overflow state within a preset sampling period when an overflow is determined to exist, and to generate an overflow trend based on the flow fluctuation rate, and to obtain the overflow development state by comparing the overflow trends of each preset sampling period. The response unit is used to match and execute corresponding response plans based on the real-time generated overflow development status.

14. The system according to claim 13, characterized in that, The reflux sampling device includes: The main body of the device includes a flow measurement chamber and multiple flow sensors mounted on the main body of the device; The flow measurement chamber includes an inlet chamber and an outlet chamber; The main body of the device introduces the recovered drilling fluid through the drilling fluid inlet; A quantitative sampler and a flow sensor A are installed on the pipeline between the drilling fluid inlet and the inlet chamber of the flow measurement chamber; Other flow sensors are installed at different locations within the flow measurement chamber to monitor drilling fluid flow at different locations.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the drilling early overflow monitoring method as described in any one of claims 1-12.

Images