AI-based intelligent control system for marine precision pressure control drilling

Abstract

The present application belongs to the field of offshore drilling, and relates to a data analysis technology, which is used to solve the problem that the prior art cannot reasonably set the control parameters when the setting value analysis results of the same control parameter are contradictory, and specifically is an offshore fine pressure control drilling intelligent control system based on AI technology, which comprises a parameter control module, a compensation control module, a protection adjustment module and a protection optimization module connected in sequence; the parameter control module is used for parameter control analysis of the seabed drilling equipment; a plurality of control groups are generated and the deviation coefficient of the control groups is calculated, and the setting optimization value is marked through the deviation coefficient; and the setting optimization value is sent to the compensation control module; the present application effectively solves the control instruction conflict problem caused by multi-source monitoring parameters, avoids the wellbore plugging risk caused by control lag, and improves the safety and control precision of deep sea drilling operations.

Description

Technical Field

[0001] This invention belongs to the field of marine drilling and involves data analysis technology, specifically an intelligent control system for marine precision pressure control drilling based on AI technology. Background Technology

[0002] The marine precision pressure controlled drilling intelligent control system is a high-end technology system used for drilling operations in deep sea or complex formations. It aims to ensure drilling safety, improve efficiency, and reduce risks by monitoring and dynamically controlling wellbore pressure in real time. The marine precision pressure controlled drilling intelligent control system is a key technology for deep-water oil and gas development, and its level of intelligence and automation directly determines the safety of operations.

[0003] The invention patent with publication number CN115822550B discloses an intelligent control system for marine precision pressure controlled drilling. This control system helps to reduce the difficulty of data acquisition and the workload of technicians in the process of marine precision pressure controlled drilling, realize the effective use of multiple parameters, and improve the accuracy of control commands and operational efficiency. However, the seabed drilling environment is complex, and the setting values ​​of the same control parameter may vary greatly from different perspectives. Under such circumstances, the existing technology cannot rationally set the control parameters, resulting in high risks in seabed drilling.

[0004] To address the aforementioned technical problems, this application proposes a solution. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control system for marine precision pressure control drilling based on AI technology, which solves the problem that existing technologies cannot rationally set control parameters when the numerical analysis results of the same control parameter settings contradict each other.

[0006] The technical problem to be solved by this invention is: how to provide an AI-based intelligent control system for marine precision pressure control drilling that can rationally set control parameters when the numerical analysis results of the same control parameter settings contradict each other.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] The intelligent control system for marine precision pressure control drilling based on AI technology includes a parameter control module, a compensation control module, a protection adjustment module, and a protection optimization module connected in sequence.

[0009] The parameter control module is used to perform parameter control analysis on the subsea drilling equipment: generating several control groups and calculating the deviation coefficients of the control groups, marking the set optimization values ​​through the deviation coefficients, and sending the set optimization values ​​to the compensation control module;

[0010] The compensation control module is used to perform calibration compensation control analysis on the subsea drilling equipment: before controlling the control parameters, compensation optimization values ​​are generated by setting optimization values, and the control parameters are numerically set by the compensation optimization values;

[0011] The protection and adjustment module is used to adjust and analyze the critical point of hydrate formation in subsea drilling equipment: it collects temperature data, fluid data, and external data from the subsea drilling equipment, inputs the collected temperature data, fluid data, and external data into an LSTM neural network model to analyze the critical point of hydrate formation and obtain the critical point threshold; if the real-time parameters under the current working conditions of the wellbore reach the critical point threshold, the current time is marked as the injection time, and the concentration and injection volume of the protective fluid are adjusted at the injection time;

[0012] The protection optimization module is used to optimize and analyze the protective fluid injection process of subsea drilling equipment.

[0013] Furthermore, the process of generating the control group includes: labeling the control parameters of the subsea drilling equipment as control objects i, i = 1, 2, ..., n, where n is a positive integer; control objects i include throttle valve opening, drilling fluid density pump, desulfurizer dosage, etc.; labeling the environmental parameters of the marine drilling as monitoring objects e, e = 1, 2, ..., m, where m is a positive integer; monitoring objects include wellbore temperature gradient, ocean current velocity, hydrogen sulfide concentration, bottom hole pressure fluctuation, etc.; each control object i corresponds to one or more monitoring objects e, and control objects i and their corresponding monitoring objects e constitute a control group. It can be understood that when control object i is the throttle valve opening, the corresponding monitoring object e corresponds to the wellbore temperature gradient (high temperature conductivity). When drilling fluid viscosity decreases, the choke valve opening needs to be increased to compensate for pressure loss. Other factors include ocean current velocity (when strong ocean currents exacerbate riser vibration, triggering choke valve fine-tuning to suppress pressure fluctuations), hydrogen sulfide concentration (when the concentration exceeds the standard, the self-drilling fluid circulation rate), and bottom hole pressure fluctuation (when the pressure deviates from the threshold by more than ±0.2MPa, the choke valve responds within 10s, combined with the drilling fluid density pump for coordinated adjustment). These factors all correspond to the choke valve opening. Therefore, the choke valve opening, wellbore temperature gradient, ocean current velocity, hydrogen sulfide concentration, and bottom hole pressure fluctuation constitute a control group. Within this control group, the wellbore temperature gradient, ocean current velocity, hydrogen sulfide concentration, and bottom hole pressure fluctuation all correspond to a theoretical setting value for the choke valve opening.

[0014] Furthermore, the process of obtaining the deviation coefficient of the control group includes: arranging the monitoring objects e within the control group in order of response priority from first to last to obtain a priority sequence. The response priority represents the priority of the control timeliness of the corresponding parameter. For example, the priority sequence of the monitoring object e within the control group corresponding to the throttle valve opening is hydrogen sulfide concentration - bottom hole pressure fluctuation - wellbore temperature gradient - ocean current velocity. The setting value SZie of the control object i is obtained by analyzing the real-time monitoring values ​​of the monitoring object e. Then, the variance of the setting values ​​SZie of all monitoring objects e corresponding to the control object i is calculated to obtain the deviation coefficient.

[0015] Furthermore, the specific process of marking the set optimization value includes: comparing the deviation coefficient with the preset deviation threshold; if the deviation coefficient is less than the deviation threshold, the average value of all the set values ​​SZie is summed to obtain the set optimization value; if the deviation coefficient is greater than or equal to the deviation threshold, the set values ​​SZie are set to the set optimization value in order of priority sequence, and the set values ​​SZie that are ranked later need to be set after the previous monitored object e has stabilized.

[0016] Furthermore, the process of generating the compensation optimization value includes: obtaining the displayed value and the actual value of the controlled object i, marking the difference between the displayed value and the actual value as the compensation value of the controlled object i, and marking the sum of the setting optimization value and the compensation value of the controlled object i as the compensation optimization value of the controlled object i.

[0017] Furthermore, the temperature data acquisition process includes: obtaining the temperature and pressure values ​​of each section of the wellbore and pipeline through temperature sensors; the fluid data acquisition process includes: detecting the gas concentration, water content, and supersaturation in the fluid composition; and the external data acquisition process includes: obtaining the seawater temperature and flow rate in the area where the wellbore is located.

[0018] Furthermore, the specific process of the protection optimization module to optimize the injection process of the protective fluid for subsea drilling equipment includes: generating an optimized conversion ratio; after the subsequent LSTM neural network model outputs the critical point threshold, marking the product of the critical point threshold and the optimized conversion ratio as the critical point optimized value; comparing the real-time parameters of the current working condition of the wellbore with the critical point optimized value and marking the subsequent injection time.

[0019] Furthermore, the process of generating the optimized conversion ratio includes: marking the time difference between the effective time of the protective fluid concentration and injection volume adjustment and the injection time as the time difference value; marking the injection optimization time, the injection optimization time being before the injection time and the time difference between the injection time and the injection time also being the time difference value; and marking the ratio of the real-time parameter at the injection optimization time to the critical point threshold as the optimized conversion ratio.

[0020] The present invention has the following beneficial effects:

[0021] 1. Effectively solves the problem of control command conflicts caused by multi-source monitoring parameters. By dynamically selecting parameter optimization strategies through deviation coefficients, it ensures the rationality and consistency of control commands. At the same time, it realizes real-time prediction of the critical point of hydrate formation and prediction and optimization of the timing of protective fluid injection, avoiding the risk of wellbore blockage caused by control lag, and improving the safety and control accuracy of deep-sea drilling operations.

[0022] 2. To address the problem of unreasonable settings caused by a single perspective in existing technologies for control parameters, by establishing a dynamic correlation between control parameters and environmental parameters, the differences in parameter settings under different analysis dimensions are reduced, thereby reducing the risk of blowouts or equipment damage caused by parameter deviations in subsea drilling operations.

[0023] 3. It can more accurately identify the risk of hydrate formation and trigger the protective fluid injection mechanism in advance when the wellbore operating conditions are close to the critical point, avoiding the problem of delayed response caused by incomplete parameter monitoring, thereby effectively reducing the potential risk of hydrate blockage of wellbore or pipeline;

[0024] 4. Effectively solves the problem of delayed injection of protective fluid. By dynamically correcting the critical point threshold and combining it with a time difference compensation mechanism, it ensures that the protective fluid is adjusted before the wellbore conditions reach the actual dangerous threshold, significantly reducing the risk of hydrate formation and improving the timeliness of injection control. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0027] Figure 2 This is a flowchart of the method in Embodiment 2 of the present invention. Detailed Implementation

[0028] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] In existing technologies, intelligent control systems for precision pressure-controlled drilling in marine environments are crucial for deepwater oil and gas development. These systems ensure drilling safety through real-time monitoring and dynamic adjustment of wellbore pressure. However, the subsea drilling environment is complex, and the set values ​​of the same control parameter obtained from different perspectives can vary significantly. Existing technologies cannot rationally set control parameters under these conditions, leading to high risks in subsea drilling. For example, in the high-pressure, low-temperature environment of the deep sea, wellbore pressure control parameters may produce contradictory analytical results due to factors such as temperature gradients and differences in fluid composition. Traditional systems struggle to reconcile conflicts between multi-source data.

[0030] To address the aforementioned issues, a control method capable of coordinating multi-dimensional parameter conflicts and achieving dynamic optimization is needed. Analysis revealed that parameter setting conflicts stem from the differentiated influence of different monitored objects on the same control parameter, necessitating the establishment of a priority ranking mechanism to eliminate these contradictions. Simultaneously, the dynamic changes in the hydrate formation critical point require the system to possess real-time predictive capabilities, necessitating a comprehensive judgment combining environmental parameters and fluid state. Furthermore, the lag in the timing of protective fluid injection may pose safety hazards, requiring the establishment of a predictive mechanism to optimize injection timing.

[0031] Example 1: As Figure 1 As shown, this application proposes an intelligent control system for marine precision pressure-controlled drilling based on AI technology, including a parameter control module, a compensation control module, a protection adjustment module, and a protection optimization module connected in sequence. The parameter control module generates control groups and calculates deviation coefficients, and sets optimized values ​​through deviation coefficient markers; the compensation control module generates compensation optimization values ​​for parameter setting; the protection adjustment module collects multi-dimensional data and inputs it into an LSTM model to analyze critical point thresholds, adjusting the protective fluid parameters when real-time parameters reach the thresholds; the protection optimization module generates an optimized conversion ratio to adjust the critical point thresholds.

[0032] The deviation coefficient of the parameter control module refers to the variance of the set values ​​for multiple monitored objects corresponding to the controlled object. This can be calculated using a variance formula and is used to quantify the comprehensive influence of different monitoring parameters on the same controlled object. The compensation optimization value of the compensation control module is the sum of the set optimization value and the deviation of the displayed value. This can be calculated by the difference between the actual values ​​collected by the sensors and the values ​​fed back by the control system, and is used to eliminate equipment execution errors. The critical point threshold of the protection adjustment module refers to the boundary value of hydrate formation conditions predicted by an LSTM neural network. This can be achieved by training a model using multi-dimensional data such as temperature, pressure, and fluid composition, and is used to dynamically determine the trigger conditions for protective fluid injection. The optimization conversion ratio of the protection optimization module is the ratio of the predicted parameter at the injection moment to the critical point threshold. This can be achieved by modeling historical effective time difference data, and is used to trigger the protective fluid injection action in advance.

[0033] Specifically, the parameter control module associates control parameters with environmental parameters to form a control group, and evaluates the dispersion of parameter settings through variance calculation. When the deviation coefficient is below the threshold, the average value is used as the optimized value; when it is above the threshold, it is set step by step according to the priority sequence. The compensation control module superimposes the equipment execution error compensation value to form the final control command. The protection adjustment module collects wellbore temperature, fluid composition, and seawater environment data in real time, inputs them into a trained LSTM model to predict the critical point, and immediately triggers the adjustment of the protective fluid concentration when the real-time pressure or temperature reaches the predicted threshold. The protection optimization module establishes an optimized conversion ratio based on the historical effective time difference, and adjusts the critical point trigger value in advance in subsequent predictions to synchronize the protective fluid injection action with changes in operating conditions.

[0034] Compared to existing technologies, current systems use a simple weighted average method to handle multi-source parameter conflicts, failing to identify the priority relationships between parameters and easily leading to conflicting control commands. This solution identifies the degree of parameter dispersion through deviation coefficients and establishes a hierarchical setting mechanism to effectively coordinate the mutual influence of multi-dimensional parameters. Existing technologies rely on fixed thresholds to determine the timing of protective fluid injection, making it difficult to adapt to dynamic environmental changes. This solution uses an LSTM model to predict the critical point in real time, combined with optimized conversion ratios to achieve injection timing prediction, significantly improving control timeliness.

[0035] Through the above technical solution, this application can effectively solve the problem of control command conflicts caused by multi-source monitoring parameters. By dynamically selecting parameter optimization strategies based on deviation coefficients, the rationality and consistency of control commands are ensured. Simultaneously, it enables real-time prediction of hydrate formation critical points and predictive optimization of protective fluid injection timing, avoiding wellbore blockage risks caused by control lags and improving the safety and control accuracy of deep-sea drilling operations.

[0036] The control parameters of the subsea drilling equipment are labeled as control objects i, i = 1, 2, ..., n, where n is a positive integer. The environmental parameters of the marine drilling are labeled as monitoring objects e, e = 1, 2, ..., m, where m is a positive integer. Each control object i corresponds to one or more monitoring objects e, and control objects i and their corresponding monitoring objects e constitute a control group.

[0037] In this system, controlled object i refers to parameters in the subsea drilling equipment that require dynamic adjustment, such as wellbore pressure, drilling fluid flow rate, or drill bit rotation speed. This can be achieved by real-time data acquisition using pressure sensors, flow meters, or rotation speed sensors. Monitored object e refers to environmental parameters related to drilling safety, such as seawater temperature, flow velocity, or formation pressure. This can be monitored using temperature sensors, flow meters, or pressure sensors. The control group consists of controlled object i and its associated monitored object e, and is used to establish a dynamic correlation between control parameters and environmental parameters, thereby avoiding the limitations of single-parameter analysis.

[0038] Specifically, when generating control groups, the control parameters are first categorized into independent control objects based on their functional types. For example, wellbore pressure is classified as one control object i, and drilling fluid flow rate as another control object i. Then, based on the actual needs of the drilling operation, each control object i is assigned associated environmental monitoring parameters. For instance, wellbore pressure control object i can be associated with monitoring objects e such as seawater temperature and flow velocity. This grouping method allows the adjustment of control parameters to be linked to real-time environmental changes. For example, when seawater flow velocity suddenly increases, the associated wellbore pressure control parameter is adjusted first, thereby achieving dynamic optimization of parameter settings.

[0039] Compared to existing technologies, the setting of control parameters in existing technologies is usually based on only single-dimensional data, such as adjusting only based on historical wellbore pressure data, while ignoring the real-time impact of environmental parameters on the control effect. This solution constructs a control group to associate environmental monitoring parameters with equipment control parameters, enabling parameter settings to comprehensively consider multi-dimensional dynamic changes. For example, when seawater temperature fluctuates, the associated drilling fluid flow rate parameter can be adjusted synchronously, thereby avoiding control deviations caused by sudden environmental changes.

[0040] Through the above technical solution, this application can solve the problem of unreasonable settings caused by the single angle of control parameter setting in the prior art. By establishing a dynamic correlation between control parameters and environmental parameters, the difference in parameter settings under different analysis dimensions can be reduced, thereby reducing the risk of blowouts or equipment damage caused by parameter deviations in subsea drilling operations.

[0041] The process of obtaining the deviation coefficient of the control group includes arranging the monitoring objects e in the control group in order of response priority to obtain the priority sequence, analyzing the real-time monitoring values ​​of the monitoring objects e to obtain the setting value SZie of the control object i, and then calculating the variance of the setting values ​​SZie of all monitoring objects e corresponding to the control object i to obtain the deviation coefficient.

[0042] Here, the priority sequence refers to the order in which monitored objects e are arranged according to their response priority. This can be achieved using preset priority rules or algorithms that dynamically adjust based on environmental parameters, and is used to determine the processing order of the influence of different monitored objects on control parameters. The setpoint value SZie refers to the parameter setpoint value corresponding to the real-time data of monitored object i under the conditions of monitored object e. This can be calculated by collecting real-time monitoring values ​​from sensors and combining them with preset mapping relationships or regression models. The deviation coefficient is a quantitative indicator reflecting the degree of dispersion of the setpoint value of the same controlled object i under different monitored objects e, calculated using variance. Specifically, it can be calculated using the variance formula in statistics to determine the numerical dispersion of multiple SZies.

[0043] Specifically, after generating the priority sequence, real-time monitoring data for each monitoring object e is input into the analysis model and converted into the corresponding setpoint value SZie for the control object i through a preset algorithm or mapping relationship. Subsequently, for the same control object i, multiple SZies are extracted and variances are calculated, with the variance results defined as the deviation coefficient. For example, when the control object is wellbore pressure, the monitoring objects may include seabed flow velocity, temperature, and fluid composition. After each monitoring object's corresponding SZie is generated through independent analysis, the variance calculation can reflect the dispersion of different monitoring objects to the pressure setpoint.

[0044] Compared to existing technologies, which fail to consider the numerical dispersion of the same control parameter across different monitoring dimensions, leading to potential conflicts in parameter settings, this solution quantifies the degree of dispersion through variance calculation and determines the processing order using a priority sequence. This ensures that the parameter setting process prioritizes data from high-priority monitoring objects, while simultaneously assessing the rationality of parameter settings through deviation coefficients.

[0045] The specific process of marking the set optimization value includes: comparing the deviation coefficient with the preset deviation threshold; if the deviation coefficient is less than the deviation threshold, the average value of all the set values ​​is obtained by summing them; if the deviation coefficient is greater than or equal to the deviation threshold, the set values ​​are set to the set optimization value in order of priority sequence, and the set values ​​that are ranked later need to be set after the previous monitored object has stabilized.

[0046] The deviation coefficient is a quantitative indicator that reflects the degree of fluctuation in the control parameter settings. Specifically, it can be achieved by calculating the variance of the corresponding settings for each monitored object, and is used to assess the stability of the control parameter settings.

[0047] The deviation threshold refers to the baseline value for determining whether the preset judgment parameter settings are stable. Specifically, it can be achieved by using the critical value obtained from the statistical analysis of historical operation data, and is used to classify the applicable conditions of parameter optimization strategies.

[0048] Priority sequence refers to the order of operations formed by sorting the monitored objects according to their response priority. Specifically, it can be achieved by setting weight coefficients for different monitoring parameters to determine the order of parameter adjustments.

[0049] Specifically, when the deviation coefficient of the control group is lower than the preset threshold, it indicates that the fluctuation of the setting values ​​corresponding to each monitored object is small. At this time, the average value of each setting value is used as the optimization value to achieve balanced parameter settings. When the deviation coefficient exceeds the threshold, it indicates that there is significant parameter fluctuation. At this time, the parameters are set step by step according to the priority order of the monitored objects. The adjustment of subsequent parameters must wait until the operating status of the previous monitored object is stable before it is executed, thereby avoiding system oscillations caused by simultaneous adjustment of multiple parameters.

[0050] Compared to existing technologies, which lack an effective mechanism for integrating different analysis results for the same control parameter, leading to the risk of conflicting parameter settings, this solution quantifies and assesses the degree of parameter fluctuation through deviation coefficients. Based on the assessment results, it dynamically selects either average optimization or step-by-step optimization strategies, ensuring both the efficiency of parameter setting in low-fluctuation scenarios and the stability of adjustment in high-fluctuation scenarios.

[0051] The process of generating the compensation optimization value includes: obtaining the displayed value and the actual value of the controlled object i, marking the difference between the displayed value and the actual value as the compensation value of the controlled object i, and marking the sum of the setting optimization value and the compensation value of the controlled object i as the compensation optimization value of the controlled object i.

[0052] Among them, the displayed value refers to the data displayed by the controlled object on the equipment operation interface or sensor. Specifically, it can be achieved by using real-time monitoring equipment such as pressure sensors and flow meters to collect and transmit the values ​​to the control system.

[0053] The actual values ​​refer to the parameter states of the controlled object in the physical environment, which can be verified through secondary verification using independent calibration equipment or redundant measuring devices.

[0054] The compensation value refers to the amount of error between the displayed value and the actual value. Specifically, it can be achieved by comparing the two in real time and generating correction parameters through the difference calculation module.

[0055] The compensation optimization value refers to the final set value of the control parameter after error correction. Specifically, it can be achieved by adding the set optimization value and the compensation value together using an adder.

[0056] Specifically, during the control parameter setting process, the displayed values ​​may deviate due to sensor drift, environmental interference, or transmission delays. By acquiring the actual values ​​and calculating compensation values, the impact of system errors on control accuracy can be eliminated. The optimized setting value is based on the initial control parameters generated from environmental parameter analysis. The compensated optimized value, obtained by superimposing the compensation value, will be output as the final execution command to the subsea drilling equipment. For example, when the displayed value of the pressure control parameter is 10 MPa while the actual value is 9.8 MPa, a compensation value of 0.2 MPa will be added to the optimized setting value to ensure that the control command matches the actual operating conditions.

[0057] Compared to existing technologies, which rely solely on a single data source for parameter setting and fail to consider systematic deviations between displayed and actual values, this application introduces a compensation value calculation mechanism to dynamically correct control parameter setting errors. For example, in patent CN115822550B, control parameters are directly generated from monitoring data without error compensation. This application, however, effectively avoids control inaccuracies caused by equipment errors through dual verification of displayed and actual values.

[0058] Through the above technical solution, this application solves the drilling risk problem caused by control parameter setting deviations in the prior art. By generating compensation optimization values ​​in real time, it is possible to ensure that the control parameters are accurately matched with the actual working conditions. For example, in wellbore pressure control, the application of compensation optimization values ​​can reduce the pressure fluctuation range to within a safe threshold, thereby avoiding hydrate formation or blowout accidents caused by parameter setting errors.

[0059] The temperature data acquisition process includes: obtaining temperature and pressure values ​​for each section of the wellbore and pipeline using temperature sensors; the fluid data acquisition process includes: detecting gas concentration, water content, and supersaturation in the fluid composition; and the external data acquisition process includes: obtaining seawater temperature and flow velocity in the area where the wellbore is located. Specifically, temperature data acquisition uses a distributed sensor network to obtain the temperature and pressure distribution of each section of the wellbore, forming a thermodynamic state profile; fluid data acquisition uses multiple types of sensors to simultaneously detect gas concentration, water content, and supersaturation in the fluid, reflecting the phase stability of the fluid inside the wellbore in real time; and external data acquisition uses marine environmental monitoring equipment to obtain seawater temperature and flow velocity around the wellbore, used to assess the impact of the external environment on the heat exchange of the wellbore. These three types of data are integrated and input into an LSTM neural network model, which uses time-series feature learning to establish a dynamic prediction model for the critical point of hydrate formation. When the real-time parameters reach the critical point threshold, the system triggers the logic for determining the injection time of the protective fluid and dynamically adjusts the concentration and injection volume of the protective fluid according to the fluid composition and environmental parameters.

[0060] The specific process of the protection optimization module in optimizing the protective fluid injection process of subsea drilling equipment includes generating an optimized conversion ratio. After the subsequent LSTM neural network model outputs a critical point threshold, the product of the critical point threshold and the optimized conversion ratio is marked as the critical point optimization value. The real-time parameters of the current wellbore operating conditions are compared with the critical point optimization value, and subsequent injection times are marked. The process of generating the optimized conversion ratio includes marking the time difference between the effective time of the adjusted protective fluid concentration and injection volume and the injection time as a time difference value, marking the injection optimization time, and marking the injection optimization time as a time difference value. The ratio of the real-time parameters at the injection optimization time to the critical point threshold is marked as the optimized conversion ratio.

[0061] The optimized conversion ratio refers to an adjustment coefficient established through the dynamic relationship between the historical adjustment effective time difference and the critical threshold. Specifically, it can be calculated using the ratio of the time difference to the real-time parameter, and is used to introduce a dynamic correction factor based on the critical threshold. The optimized critical threshold value refers to the adjustment result of the original critical threshold based on the optimized conversion ratio, which can be achieved through multiplication operations, and is used to trigger the protective fluid injection action in advance. The optimized injection time refers to the pre-set start time point for protective fluid injection based on the effective time difference, which can be calculated by shifting the time difference forward, and is used to reserve the response period required for the protective fluid to take effect.

[0062] Specifically, after the LSTM neural network model outputs the critical threshold for hydrate formation, the optimized conversion ratio is dynamically calculated and applied to the optimized value for the critical threshold. Real-time wellbore parameters are continuously compared with the optimized critical threshold value. If the real-time parameters reach this optimized value, an injection time marker is triggered. At this point, the protective fluid injection action is advanced to the optimized injection time, ensuring that the concentration and injection volume of the protective fluid are adjusted before the actual critical condition is reached, thereby avoiding the risk of hydrate formation due to delayed activation.

[0063] The process of generating the optimized conversion ratio includes marking the time difference between the effective time of the adjusted protective fluid concentration and injection volume and the injection time as the time difference value, marking the injection optimization time, the injection optimization time being before the injection time and the time difference between the injection time and the injection time also being the time difference value, and marking the ratio of the real-time parameters at the injection optimization time to the critical point threshold as the optimized conversion ratio.

[0064] The time difference refers to the time interval required from the execution of the protective fluid concentration or injection volume adjustment operation to its actual effect. Specifically, it can be achieved by using a timer module to record the time difference between the execution time of the adjustment command and the time when the sensor feedback parameters stabilize. This feature is used to quantify the response delay characteristics of the protective fluid adjustment operation.

[0065] Among them, the injection optimization time refers to the optimal start time of the protective fluid adjustment operation calculated based on the time difference value. Specifically, it can be determined by shifting the time difference value forward based on the injection time. This feature is used to establish a dynamic time correlation between the protective fluid adjustment operation and the critical point threshold.

[0066] Among them, the optimization conversion ratio refers to the dynamic proportional coefficient between the real-time operating parameters at the injection optimization time and the critical point threshold. Specifically, it can be achieved by calculating the ratio between the real-time collected wellbore temperature and pressure data and the critical point threshold output by the LSTM model. This feature is used to construct a dynamic matching mechanism between the timing of protective fluid injection and changes in operating parameters.

[0067] Specifically, during the protective fluid injection optimization process, the time difference between the execution and effectiveness of protective fluid concentration or injection volume adjustment operations is first recorded using historical data as a time difference value. Once the LSTM neural network model outputs the critical threshold for the current operating condition, the system automatically shifts the injection time forward by the time difference value, marking it as the injection optimization time when the protective fluid adjustment operation needs to be initiated. At this time, wellbore temperature and pressure parameters are collected in real-time at the injection optimization time, and their ratio is calculated with the critical threshold to generate an optimization conversion ratio. This conversion ratio is applied to the subsequent dynamic correction of the critical threshold, enabling the protective fluid adjustment operation to be prepared for effectiveness before the actual operating conditions reach the critical threshold.

[0068] Example 2: Figure 2 As shown, the intelligent control method for marine precision pressure-controlled drilling based on AI technology includes the following steps:

[0069] Step 1: Perform parameter control analysis on the subsea drilling equipment: A control group is formed by the control object i and the corresponding monitoring object e. The variance of the set value SZie of all monitoring objects e corresponding to the control object i in the control group is calculated to obtain the deviation coefficient. The value of the control object i is controlled by the deviation coefficient.

[0070] Step 2: Calibration and compensation control analysis of the subsea drilling equipment: Obtain the compensation optimization value by comparing the displayed value and the actual value of the controlled object, and use the compensation optimization value to set the parameters of the controlled object i;

[0071] Step 3: Adjustment analysis of the critical point for hydrate formation of subsea drilling equipment: Temperature data, fluid data, and external data are collected from the subsea drilling equipment. The collected temperature data, fluid data, and external data are input into an LSTM neural network model to analyze the critical point for hydrate formation and obtain the critical point threshold. The injection time is marked using the critical point threshold.

[0072] Step 4: Optimize and analyze the protective fluid injection process of the subsea drilling equipment: Mark the time difference between the effective time of the adjusted protective fluid concentration and injection volume and the injection time as the time difference value, and mark the critical point optimization value through the time difference value.

[0073] The intelligent control system for marine precision pressure-controlled drilling based on AI technology consists of the following modules: During operation, the parameter control module generates control groups and calculates deviation coefficients, and sets optimized values ​​by marking deviation coefficients; the compensation control module generates compensation optimized values ​​for parameter setting; the protection adjustment module collects multi-dimensional data and inputs it into an LSTM model to analyze critical point thresholds, and adjusts protective fluid parameters when real-time parameters reach the thresholds; and the protection optimization module generates an optimized conversion ratio to adjust the critical point thresholds.

[0074] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

[0075] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0076] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent control system for fine pressure control drilling at sea based on AI technology, characterized in that, The parameter control module, the compensation control module, the protection adjustment module and the protection optimization module are sequentially connected. The parameter control module is configured to perform parameter control analysis on the subsea drilling equipment, generate a plurality of control groups, calculate deviation coefficients of the control groups, and mark the setting optimization values by using the deviation coefficients; and send the setting optimization values to the compensation control module. The compensation control module is configured to perform calibration compensation control analysis on the subsea drilling equipment, generate compensation optimization values by using the setting optimization values before controlling the control parameters, and set the control parameters by using the compensation optimization values. The protection adjustment module is configured to perform adjustment analysis on the hydrate generation critical point of the subsea drilling equipment, collect temperature data, fluid data and external data of the subsea drilling equipment, input the collected temperature data, fluid data and external data into an LSTM neural network model to analyze the hydrate generation critical point, and obtain a critical point threshold. If the real-time parameters of the wellbore under the current working condition reach the critical point threshold, the current time is marked as an injection time, and the concentration and injection amount of the protection fluid are adjusted at the injection time. The protection optimization module is configured to perform optimization analysis on the protection fluid injection process of the subsea drilling equipment. The deviation coefficient of the control group is obtained by arranging the monitoring objects e in the control group in the order of response priority to obtain a priority sequence, analyzing the setting values SZie of the control objects i according to the real-time monitoring values of the monitoring objects e, and calculating the variance of the setting values SZie of all monitoring objects e corresponding to the control object i to obtain the deviation coefficient. The setting optimization value is marked by comparing the deviation coefficient with a preset deviation threshold, summing and averaging all setting values SZie to obtain the setting optimization value if the deviation coefficient is less than the deviation threshold, and setting the setting values SZie as the setting optimization value in the order of the priority sequence if the deviation coefficient is greater than or equal to the deviation threshold. The generation process of the compensation optimization value includes obtaining the indicated value and the actual value of the control object i, marking the difference between the indicated value and the actual value as the compensation value of the control object i, and marking the sum of the setting optimization value and the compensation value of the control object i as the compensation optimization value of the control object i. 2.The AI technology-based intelligent control system for fine pressure control drilling at sea according to claim 1, characterized in that, The control group is generated by marking the control parameters of the subsea drilling equipment as control objects i, i = 1, 2, …, n, n being a positive integer, marking the environmental parameters of the subsea drilling equipment as monitoring objects e, e = 1, 2, …, m, m being a positive integer, and forming a control group by the control object i and the corresponding monitoring object e. 3.The AI technology-based intelligent control system for fine pressure control drilling at sea according to claim 2, characterized in that, The temperature data acquisition process includes obtaining the temperature and pressure values of each section of the wellbore and the pipeline by using a temperature sensor; the fluid data acquisition process includes detecting the gas concentration, water content and supersaturation in the fluid composition; and the external data acquisition process includes obtaining the seawater temperature and flow rate of the region where the wellbore is located. 4.The AI technology-based intelligent control system for fine marine pressure control drilling according to claim 3, characterized in that, The specific process of the protection optimization module for optimizing and analyzing the protection fluid injection process of the subsea drilling equipment includes: generating an optimized conversion ratio, marking the product of the critical point threshold value and the optimized conversion ratio as a critical point optimization value after the subsequent LSTM neural network model outputs the critical point threshold value, comparing the real-time parameters of the current working condition of the wellbore with the critical point optimization value and marking the subsequent injection time; The generation process of the optimized conversion ratio includes: marking the time difference between the effective time of the protection fluid concentration and the injection amount adjustment and the injection time as a time difference value, marking an injection optimization time, the injection optimization time being before the injection time and the time difference value between the injection optimization time and the injection time also being the time difference value, and marking the ratio of the real-time parameters of the injection optimization time to the critical point threshold value as the optimized conversion ratio.

Images