A casing head pressure self-balancing and annular pressure management control system

Abstract

The application provides a casing head pressure self-balancing and annular space pressure management control system, relates to the oil field oil and gas exploitation technical field and comprises a real-time sensing module, a three-dimensional pressure decoupling module, a double closed loop self-adaptive control module, a double mode execution module, an annular space protective fluid intelligent injection module, a cloud self-adaptive learning module and a control center; the application separates the coupled interference signals by a three-channel real-time sensing network and a three-dimensional pressure decoupling model in combination with a self-adaptive Kalman filtering algorithm, realizes independent identification and dynamic weight distribution of each pressure source, introduces a three-dimensional pressure decoupling mechanism in the casing head system for the first time, effectively breaks through the control instability bottleneck caused by the casing pressure-annular space pressure coupling, avoids control oscillation caused by the traditional single-point feedback, significantly improves the stability of the casing head pressure control, adapts to the complex working conditions of the coexistence of multiple pressure systems and provides a reliable pressure control basis for deep oil and gas exploitation.

Description

Technical Field

[0001] This invention relates to the field of oilfield oil and gas extraction technology, and in particular to a casing head pressure self-balancing and annular pressure management and control system. Background Technology

[0002] In oilfield oil and gas extraction, the casing head, as the core connection between the wellbore and surface equipment, directly determines the integrity of the wellbore barrier and the safety of extraction operations, making it a key component of the well control system. Currently, the industry commonly uses casing head pressure control devices, which achieve preliminary control of casing pressure and annular pressure by adjusting the opening of the balance valve group through single-point pressure sensor feedback and single closed-loop control logic. Annular protection fluid is mostly injected in a timed and quantitative manner, relying on manual or simple automatic control to adjust the injection volume. Some high-end systems introduce single-modal actuators, combined with basic control algorithms, in an attempt to improve pressure control accuracy.

[0003] However, casing pressure, annular pressure, and formation pressure are strongly coupled, and existing systems lack effective decoupling mechanisms. Single-point feedback easily leads to control oscillations, making it impossible to independently identify each pressure source, which in turn causes control instability. This is one of the core bottlenecks in current ultra-high pressure oil and gas well control. Single closed-loop control logic has lag and large overshoot, making it unable to adapt to dynamic changes in casing head pressure, and it is difficult to balance control accuracy and response speed. Execution units mostly use single modes, resulting in either insufficient response speed or insufficient adjustment accuracy, making it impossible to achieve rapid response during pressure surges and accurate maintenance during steady state, and difficult to adapt to complex pressure change scenarios. Annular protective fluid injection does not consider the coupling effect of temperature, pressure, and flow rate, and timed injection easily leads to overpressure or underpressure, affecting the integrity of the wellbore barrier. The system lacks adaptive learning capabilities and cannot optimize control strategies based on different well conditions and historical operating data, making it difficult to continuously improve control performance. Furthermore, data fusion and intelligent technology lag behind, limiting further improvement in control accuracy. Therefore, this invention proposes a casing head pressure self-balancing and annular pressure management control system to solve the problems existing in the prior art. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a casing head pressure self-balancing and annular pressure management control system. This system utilizes a three-channel real-time sensor network and a three-dimensional pressure decoupling model, combined with an adaptive Kalman filter algorithm to separate coupled interference signals. This enables independent identification and dynamic weight allocation of each pressure source. For the first time, a three-dimensional pressure decoupling mechanism is introduced into the casing head system, effectively overcoming the control instability bottleneck caused by strong coupling between casing pressure and annular pressure. It avoids control oscillations caused by traditional single-point feedback, significantly improving the stability of casing head pressure control. This system is adaptable to complex operating conditions with multiple pressure systems coexisting, providing a reliable pressure control foundation for deep oil and gas extraction.

[0005] To achieve the objectives of this invention, the following technical solution is provided: a casing head pressure self-balancing and annular pressure management and control system, comprising a real-time sensing module, a three-dimensional pressure decoupling module, a dual-closed-loop adaptive control module, a dual-modal execution module, an annular protective fluid intelligent injection module, a cloud adaptive learning module, and a control center. The real-time sensing module constructs a three-channel real-time sensing network for casing pressure, annular pressure, and formation pressure, used to collect pressure, temperature, and flow parameters related to the casing head and wellbore. The three-dimensional pressure decoupling module constructs a three-dimensional pressure decoupling model based on the collected pressure parameters, and separates coupled interference signals using an adaptive Kalman filter algorithm to achieve independent identification and dynamic weight allocation of each pressure source. The dual-closed-loop adaptive control module employs a dual-layer closed-loop logic of inner-loop adjustment + outer-loop calibration to achieve dynamic self-balancing of the casing head pressure. The dual-modal execution module executes pressure regulation actions, achieving rapid pressure response and high-precision pressure maintenance.

[0006] The intelligent injection module for annular protective fluid is connected to the real-time sensing module and the control center, and realizes dynamic injection of annular protective fluid based on a three-parameter coupling model of temperature-pressure-flow rate; the cloud adaptive learning module is connected to the control center, and realizes dynamic updating of control parameters by training a prediction model and control strategy library through machine learning algorithms; the control center coordinates the collaborative work of each module and automatically switches the control mode according to the real-time pressure gradient difference.

[0007] A further improvement lies in the following: The construction process of the three-dimensional pressure decoupling model is as follows: using casing pressure P1, annular pressure P2, and formation pressure P3 as input variables, a three-dimensional pressure coupling equation is constructed. An adaptive Kalman filter algorithm is used to separate the coupling signals, obtaining the independent signals of each pressure source, thus achieving dynamic weight allocation. The state equation of the adaptive Kalman filter algorithm is:

[0008] Xk = AXk-1 + BUk + Wk,

[0009] The observation equation is:

[0010] Zk=HXk+Vk,

[0011] Where Xk is the state vector at time k, containing the true values ​​of P1, P2, and P3; A is the state transition matrix, a 3×3 identity matrix; B is the control matrix, a 3×1 zero matrix; Uk is the control input at time k, with a value of 0; Wk is the process noise, following a Gaussian distribution with a mean of 0 and a variance of Q, where Q is a 3×3 diagonal matrix with diagonal elements of 0.01, 0.01, and 0.02; Zk is the observation vector at time k, i.e., the pressure value collected by the real-time sensing module; H is the observation matrix, a 3×3 identity matrix; Vk is the observation noise, following a Gaussian distribution with a mean of 0 and a variance of R, where R is a 3×3 diagonal matrix with diagonal elements of 0.005, 0.005, and 0.01.

[0012] Further improvements are made in the following: The control logic of the dual closed-loop adaptive control module is as follows: The inner loop uses the deviation ΔP between the real-time pressure of the bushing head and the preset pressure threshold as the control target, and adjusts the opening of the balancing valve group and the power of the compensation pump through the pressure self-balancing execution module to achieve rapid pressure response; the outer loop aims at the stability of pressure regulation, collects the pressure fluctuation data after the inner loop adjustment in real time, and dynamically corrects the inner loop control parameters through an adaptive fuzzy PID algorithm. The inner loop control parameters include the proportional coefficient Kp, integral time Ti, and derivative time Td; the parameter correction formula of the adaptive fuzzy PID algorithm is:

[0013] ΔKp=Kp0+Kp1×E+Kp2×EC,

[0014] ΔTi = Ti0 + Ti1 × E + Ti2 × EC

[0015] ΔTd = Td0 + Td1 × E + Td2 × EC;

[0016] Wherein, ΔKp, ΔTi, and ΔTd are the correction amounts for the proportional coefficient, integral time, and derivative time, respectively; Kp0, Ti0, and Td0 are the initial control parameters, with values ​​of 5.0, 1.2s, and 0.3s, respectively; Kp1, Kp2, Ti1, Ti2, Td1, and Td2 are the correction coefficients, with values ​​of 0.8, 0.4, 0.1, 0.05, 0.08, and 0.04, respectively; E is the fuzzy quantized value of the pressure deviation ΔP, and EC is the fuzzy quantized value of the pressure deviation change rate.

[0017] Further improvements include: the dual-mode execution module comprises a main channel and a secondary channel. The main channel is a high-speed servo hydraulic valve with a response time ≤50ms, used for rapid pressure relief or replenishment during sudden pressure changes; the secondary channel is a precision electrically controlled needle valve driven by a stepper motor, with a pressure regulation resolution of 0.01MPa, used for high-precision pressure maintenance in steady state; the control center automatically switches the execution mode according to the pressure change rate v: when v > 0.3MPa / s, the main channel high-speed servo hydraulic valve is activated for rapid adjustment; when v < 0.05MPa / s, it switches to the secondary channel precision electrically controlled needle valve for fine adjustment; when 0.05MPa / s ≤ v ≤ 0.3MPa / s, the two channels work together; the dual-mode execution module also has a fault redundancy module, so that when either channel fails, the other channel automatically switches to the main control channel to ensure continuous pressure regulation.

[0018] A further improvement lies in the following: the intelligent injection module for annular protection fluid is based on a real-time temperature-pressure-flow rate three-parameter coupling model to construct an annular protection fluid injection volume prediction function, which is:

[0019] Q = K × (P × T) / F;

[0020] Where Q is the annular protective fluid injection volume, in L / min; K is the injection coefficient, ranging from 0.02 to 0.05, dynamically adjusted according to well conditions; P is the real-time annular pressure, in MPa; T is the real-time annular temperature, in °C; and F is the protective fluid viscosity coefficient, in mPa·s, preset according to the protective fluid type. The system automatically adjusts the injection pump frequency and flow rate based on this prediction function to achieve second-level dynamic compensation, with an injection response time ≤1s.

[0021] Further improvements include: the cloud adaptive learning module comprises a cloud server and a local data interaction unit. The cloud server deploys an adaptive learning engine to collect historical operating data from multiple well casing heads. This historical operating data includes annular pressure variation curves, temperature variation curves, control action records, and corresponding production condition parameters. The adaptive learning engine uses a random forest algorithm to train a dynamic annular pressure behavior prediction model. The prediction formula for the random forest algorithm is:

[0022] Y = (1 / N) × ΣYi;

[0023] Where Y is the predicted annular pressure value in MPa; N is the number of decision trees, ranging from 100 to 200; Yi is the predicted value of the i-th decision tree in MPa; and the gradient boosting decision tree algorithm is used to train the optimal control strategy library to achieve adaptive matching of control parameters under different well conditions.

[0024] Further improvements include: the local data interaction unit establishes a connection with the cloud server via industrial Ethernet or 5G wireless communication technology, and periodically downloads updated prediction model parameters and control strategy parameters from the cloud server, with an update cycle of 1 to 24 hours, dynamically adjusted according to the complexity of the well conditions; simultaneously, the local data interaction unit uploads the actual effect data of each pressure regulation to the cloud server, including the pressure values ​​before and after regulation, regulation time, and overshoot, as training samples for continuous model optimization; the cloud server is also equipped with a data encryption module to ensure the security of data transmission and storage.

[0025] Further improvements include: the control center offers two control modes: self-balancing dominant and annular compensation dominant. The control center automatically switches modes based on the real-time pressure deviation value ΔP: when ΔP < 0.02 MPa / m, it switches to self-balancing dominant mode, with the dual-closed-loop adaptive control module leading pressure regulation; when ΔP ≥ 0.02 MPa / m, it switches to annular compensation dominant mode, with the annular protection fluid intelligent injection module leading pressure regulation, and the dual-closed-loop adaptive control module assisting in regulation. The control center also includes an alarm module; when the pressure parameter exceeds a preset safety threshold of ±0.2 MPa, it automatically issues an audible and visual alarm and triggers an emergency pressure relief action.

[0026] Further improvements include: the real-time sensing module includes a casing pressure sensor, annular pressure sensor, formation pressure sensor, temperature sensor, and flow sensor. The sampling frequency of each sensor is 10~50Hz, the pressure measurement accuracy is ±0.01MPa, the temperature measurement accuracy is ±0.1℃, and the flow measurement accuracy is ±0.1L / min. The sensors adopt a corrosion-resistant and high-pressure-resistant design, with an operating pressure range of 0~100MPa and an operating temperature range of -40℃~150℃, making them suitable for extreme downhole conditions in oilfields.

[0027] Further improvements include: the system also has a data storage module and a remote monitoring module. The data storage module is used to store real-time sensor data, control parameters, and fault records; the remote monitoring module allows administrators to view the system's operating status and parameter curves in real time through terminal devices, and provides the function of remotely issuing control commands to achieve remote operation and maintenance.

[0028] The beneficial effects of this invention are as follows:

[0029] 1. This invention uses a three-channel real-time sensor network and a three-dimensional pressure decoupling model, combined with an adaptive Kalman filter algorithm to separate coupled interference signals, achieving independent identification and dynamic weight allocation of each pressure source. It is the first time that a three-dimensional pressure decoupling mechanism has been introduced into the casing head system, effectively breaking through the control instability bottleneck caused by strong coupling between casing pressure and annular pressure, avoiding control oscillations caused by traditional single-point feedback, significantly improving the stability of casing head pressure control, adapting to complex working conditions with multiple pressure systems coexisting, and providing a reliable pressure control foundation for deep oil and gas extraction.

[0030] 2. This invention employs a dual-closed-loop adaptive pressure self-balancing control logic, constructing a two-layer closed-loop system of inner-loop regulation + outer-loop calibration. The inner loop achieves rapid pressure response, while the outer loop dynamically corrects the control parameters of the inner loop, eliminating regulation overshoot and oscillation phenomena. The regulation response speed is improved by more than 30% compared to traditional single-closed-loop control, and the pressure control accuracy error is controlled within ±0.1MPa. This effectively solves the technical defects of traditional control, such as regulation lag and large overshoot, ensuring the dynamic self-balancing of casing head pressure and meeting the stringent requirements for pressure control accuracy in ultra-high pressure oil and gas extraction.

[0031] 3. This invention employs a dual-mode actuator. The main channel high-speed servo hydraulic valve enables rapid response to sudden pressure changes, while the secondary channel precision electrically controlled needle valve enables high-precision fine-tuning during steady-state operation. The system automatically switches the execution mode according to the rate of pressure change. For the first time, this invention integrates high-power hydraulic response with high-precision electric adjustment capabilities in a casing head system, achieving fast-precision coordinated control and filling a gap in the industry. At the same time, a fault redundancy mechanism is added to ensure the reliability of the actuator and avoid well control risks caused by execution failures.

[0032] 4. This invention is based on a real-time temperature-pressure-flow rate three-parameter coupling model to construct an annular protective fluid injection volume prediction function. The system automatically adjusts the injection pump frequency and flow rate according to the algorithm to achieve second-level dynamic compensation, avoiding the risk of overpressure or underpressure caused by traditional timed injection. Thermodynamic parameters are incorporated into the injection control logic to achieve intelligent compensation by controlling pressure with temperature and adjusting fluid with pressure, significantly improving the integrity of the wellbore barrier, reducing the risk of casing damage and oil and gas leakage, and extending the service life of the wellbore.

[0033] 5. This invention deploys an adaptive learning engine on a cloud server to collect historical operating data from multiple oil well casing heads. It uses machine learning algorithms to train a prediction model and an optimal control strategy library, enabling data interaction and parameter updates between the local control unit and the cloud server. This allows the system to adapt to dynamic changes in different well conditions. As operating data accumulates, prediction accuracy and control effectiveness continuously improve, achieving a "the more you use it, the smarter it becomes" technical effect. At the same time, it reduces network latency through edge computing nodes, solving the problems of lagging data fusion and low intelligence level in existing systems, and reducing manual operation and maintenance costs. Attached Figure Description

[0034] Figure 1 This is a diagram illustrating the composition of the present invention. Detailed Implementation

[0035] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0036] Example 1

[0037] according to Figure 1 As shown, this embodiment proposes a casing head pressure self-balancing and annular pressure management and control system to verify the effectiveness of the three-dimensional pressure decoupling module and adapt to the working conditions of strong coupling of three pressure sources in deep oil and gas extraction. The specific implementation process is as follows:

[0038] Sensor Deployment: A casing pressure sensor is installed at the top of the casing head, an annulus pressure sensor is installed at the annulus interface, and a formation pressure sensor is installed on the inner wall of the formation casing to construct a three-channel real-time sensor network. The sensor sampling frequency is set to 30Hz, the pressure measurement accuracy is ±0.01MPa, the working pressure range is 0~80MPa, and the working temperature range is -20℃~120℃. It adopts anti-corrosion coating treatment, is suitable for complex downhole environments, and meets the sensor installation and performance requirements in the standard.

[0039] Model Construction: Based on the casing pressure P1, annular pressure P2, and formation pressure P3 data collected by the real-time sensing module, a three-dimensional pressure coupling equation is constructed. The state equation and observation equation of the adaptive Kalman filter algorithm are substituted into the equation. The process noise variance matrix Q is set as diag([0.01, 0.01, 0.02]), the observation noise variance matrix R is set as diag([0.005, 0.005, 0.01]), the state transition matrix A and the observation matrix H are both 3×3 identity matrices, the control matrix B is a 3×1 zero matrix, and the control input Uk is 0.

[0040] Test process: Simulate the pressure coupling scenario in deep oil and gas extraction, artificially apply coupling interference signals to make the coupling degree between P1, P2 and P3 reach 30%~50%, and use a three-dimensional pressure decoupling module to separate and identify the signals independently. Record the signal fluctuation of each pressure source before and after decoupling, and compare the control stability of the traditional non-decompression method with the decoupling method of this invention.

[0041] Implementation Results: After 72 hours of continuous testing, the three-dimensional pressure decoupling module of this invention can effectively separate coupled interference signals. After decoupling, the independent identification accuracy of each pressure source reaches over 98%, the pressure fluctuation amplitude is reduced by over 60%, and the control oscillation phenomenon is completely eliminated. Compared with the traditional non-decompression method, the number of control instabilities is reduced from 3-5 times per hour to 0 times. It successfully breaks through the control instability bottleneck caused by strong coupling of "casing pressure-annular air pressure", verifying the rationality and effectiveness of the module, and it can adapt to the complex pressure environment of deep oil and gas extraction.

[0042] Example 2

[0043] according to Figure 1 As shown, this embodiment proposes a casing head pressure self-balancing and annular pressure management control system to verify the control accuracy and response speed of the dual closed-loop adaptive control module and compare the performance differences with the traditional single closed-loop control logic. The specific implementation process is as follows:

[0044] System Setup: In the casing head system of a major oil well in a certain oilfield, a dual closed-loop adaptive control module was built. The inner loop is equipped with a pressure self-balancing execution module, including a balancing valve group and a compensation pump, with a preset casing head pressure threshold of 25MPa. The outer loop is equipped with a parameter correction module, which adopts an adaptive fuzzy PID algorithm. The initial control parameters are Kp0=5.0, Ti0=1.2s, Td0=0.3s, and the correction coefficients are Kp1=0.8, Kp2=0.4, Ti1=0.1, Ti2=0.05, Td1=0.08, Td2=0.04. The fuzzy universe of discourse for pressure deviation E and deviation change rate EC is (-8,8), and the fuzzy subset is divided into 9 levels.

[0045] Test conditions: Simulate dynamic pressure changes in the casing head, including three conditions: pressure surge (from 25MPa to 30MPa, rate of change 0.4MPa / s), pressure drop (from 25MPa to 20MPa, rate of change 0.5MPa / s), and steady-state fluctuation (pressure fluctuates within the range of 25MPa±0.5MPa). The dual closed-loop control logic of this invention and the traditional single closed-loop control logic are used for testing respectively, and the adjustment response time, overshoot, and control accuracy are recorded.

[0046] Implementation Results: Test results show that the dual-closed-loop adaptive control module of this invention has a response time of 0.8s, an overshoot of 0.3MPa, and a control accuracy error of ±0.08MPa under pressure surge conditions; under pressure drop conditions, the response time is 0.7s, the overshoot is 0.25MPa, and the control accuracy error is ±0.07MPa; and under steady-state fluctuation conditions, the pressure fluctuation amplitude is controlled within ±0.05MPa. Compared with the traditional single-closed-loop control logic, the response speed is improved by 35%, the overshoot is reduced by 70%, and the control accuracy is improved by 40%. It effectively solves the technical defects of traditional control, such as lag and large overshoot, fully meets the requirements of dynamic self-balancing of casing head pressure, and is adaptable to different pressure change scenarios.

[0047] Example 3

[0048] according to Figure 1 As shown, this embodiment proposes a casing head pressure self-balancing and annular pressure management control system to verify the fast-precision collaborative control effect of the dual-modal actuator and test the execution performance under different pressure change rates. The specific implementation process is as follows:

[0049] Actuation Unit Deployment: A dual-modal actuator is installed. The main channel high-speed servo hydraulic valve is an industrial-grade servo valve with a response time of 45ms, a rated working pressure of 0~100MPa, and a maximum flow rate of 50L / min. The secondary channel precision electrically controlled needle valve is driven by a stepper motor, with a pressure regulation resolution of 0.008MPa, a rated working pressure of 0~80MPa, and an adjustment range of 0~10L / min. Pressure change rate thresholds are set: v>0.3MPa / s is a pressure jump, v<0.05MPa / s is a steady-state fine adjustment, and 0.05MPa / s≤v≤0.3MPa / s is a coordinated adjustment. A fault redundancy module is also deployed, and the channel fault detection response time is set to ≤100ms.

[0050] Test Procedure: Three pressure change scenarios were simulated: Scenario 1, sudden pressure change (from 20MPa to 28MPa, rate of change 0.4MPa / s), testing the rapid response capability of the main channel; Scenario 2, steady-state fine-tuning (pressure maintained at 25MPa, fluctuation rate 0.03MPa / s), testing the high-precision adjustment capability of the secondary channel; Scenario 3, gradual pressure change (from 25MPa to 27MPa, rate of change 0.1MPa / s), testing the coordinated adjustment capability of the dual channels; at the same time, a main channel failure scenario was simulated to test the effectiveness of the fault redundancy module.

[0051] Implementation Results: In Scenario 1, the main channel high-speed servo hydraulic valve starts within 0.1s and completes pressure regulation within 1.2s, with an regulation error of ±0.1MPa, achieving rapid pressure relief and replenishment. In Scenario 2, the secondary channel precision electro-hydraulic needle valve controls pressure fluctuations within ±0.005MPa, achieving nanometer-level pressure maintenance. In Scenario 3, the dual channels work together, with a regulation response time of 0.5s and a regulation accuracy of ±0.06MPa, balancing response speed and regulation accuracy. In the event of a main channel failure, the secondary channel automatically switches to the main control channel within 0.1s, ensuring uninterrupted pressure regulation, and maintaining regulation accuracy within ±0.12MPa even in fault conditions. This embodiment verifies the fast-precision collaborative control effect of the dual-modal actuator, filling an industry gap, and the fault redundancy mechanism improves system reliability.

[0052] Example 4

[0053] according to Figure 1 As shown, this embodiment proposes a casing head pressure self-balancing and annular pressure management and control system to verify the dynamic compensation effect of the intelligent annular protection fluid injection module, avoid overpressure or underpressure risks, and improve wellbore barrier integrity. The specific implementation process is as follows:

[0054] Module Deployment: Flow sensors and temperature sensors are installed on the annular protective fluid injection pipeline, working in conjunction with the annular pressure sensor to construct a three-parameter acquisition network of temperature, pressure, and flow rate; the injection pump is a variable frequency speed control pump with an adjustment range of 0~20L / min and a frequency adjustment accuracy of 0.1Hz; a prediction function for the annular protective fluid injection volume is constructed as Q = K×(P×T) / F, where K is set to 0.03, F is preset to 50mPa·s based on the protective fluid model, and the injection response time is set to 0.8s.

[0055] Test conditions: Simulate different well conditions with varying annular pressure and temperature, including high temperature and high pressure (80℃, 35MPa), normal temperature and pressure (25℃, 15MPa), and low temperature and low pressure (10℃, 10MPa). Tests were conducted using the intelligent injection method of this invention and the traditional timed injection method (100L per hour), respectively, and annular pressure fluctuations, injection volume errors, and wellbore sealing performance were recorded.

[0056] Implementation Results: Test results show that under high temperature and high pressure conditions, the intelligent injection module of this invention has an injection volume error of ±0.5L / min, annular pressure fluctuation controlled within ±0.08MPa, and good wellbore sealing performance with no leakage. Under normal temperature and pressure conditions, the injection volume error is ±0.3L / min, and annular pressure fluctuation controlled within ±0.05MPa. Under low temperature and low pressure conditions, the injection volume error is ±0.4L / min, and annular pressure fluctuation controlled within ±0.06MPa. Traditional timed injection methods, under high temperature and high pressure conditions, have an injection volume error of ±3L / min and annular pressure fluctuation of ±0.5MPa, exhibiting slight overpressure; under low temperature and low pressure conditions, underpressure occurs, and wellbore sealing performance deteriorates. This embodiment verifies that the intelligent injection module can achieve second-level dynamic compensation, avoid overpressure or underpressure risks, significantly improve wellbore barrier integrity, and adapt to different temperature and pressure conditions.

[0057] Example 5

[0058] according to Figure 1 As shown, this embodiment proposes a casing head pressure self-balancing and annular pressure management and control system to verify the self-optimization effect of the cloud adaptive learning module and achieve adaptive matching for different well conditions. The specific implementation process is as follows:

[0059] System Setup: A cloud server was deployed, an adaptive learning engine was installed, and historical operating data of the casing head from 100 oil wells with different well conditions was collected, including annular pressure change curves, temperature change curves, control action records, and production parameters (well depth, production rate, formation permeability), totaling 10 million data entries. A random forest algorithm was used to train a dynamic behavior prediction model for annular pressure, with N=150 decision trees. A gradient boosting decision tree algorithm was used to train the optimal control strategy library, with 1000 training iterations. The local control unit was connected to the cloud server via 5G wireless communication, with a parameter update cycle of 12 hours and a data upload cycle of 1 hour. A data encryption module was deployed, using the AES-256 encryption algorithm to ensure data security.

[0060] Test process: Ten oil wells with different conditions (including deep wells, shallow wells, high-permeability wells, and low-permeability wells) were selected. The system of this invention was applied to these oil wells, and the initial control effect and the control effect after 1 month, 3 months, and 6 months of cloud learning optimization were recorded. The changes in prediction accuracy, adjustment response speed, and control accuracy were compared. At the same time, a new well condition (abrupt change in formation permeability) was simulated to test the adaptive matching capability of the system.

[0061] Implementation Results: After six months of cloud-based learning and optimization, the system's annular pressure prediction accuracy improved from the initial 88% to 98.5%, the adjustment response speed increased by 15%, and the control accuracy error decreased from ±0.08MPa to ±0.05MPa. In new well condition simulations, the system downloaded updated control strategy parameters from the cloud server and completed adaptive matching within 0.5 hours, demonstrating stable control performance. The adjustment response time and overshoot both met design requirements. Compared to systems without cloud learning, the system's adaptive capability is significantly improved, enabling it to adapt to dynamic changes in different well conditions and achieve a "the more it's used, the smarter it becomes" technical effect. Simultaneously, it reduces manual maintenance costs, decreasing monthly maintenance workload by more than 30% per well.

[0062] This casing head pressure self-balancing and annular pressure management and control system utilizes a three-channel real-time sensor network and a three-dimensional pressure decoupling model, combined with an adaptive Kalman filter algorithm to separate coupled interference signals. This enables independent identification and dynamic weight allocation of each pressure source. For the first time, a three-dimensional pressure decoupling mechanism is introduced into the casing head system, effectively overcoming the control instability bottleneck caused by strong coupling between casing pressure and annular pressure, avoiding control oscillations caused by traditional single-point feedback, significantly improving the stability of casing head pressure control, and adapting to complex operating conditions with multiple pressure systems coexisting. This provides a reliable pressure control foundation for deep oil and gas extraction. The invention employs a dual-closed-loop adaptive pressure self-balancing control logic, constructing a two-layer closed-loop system of inner-loop regulation + outer-loop calibration. The inner loop achieves rapid pressure response, while the outer loop dynamically corrects the inner loop control parameters, eliminating overshoot and oscillation phenomena. The regulation response speed is improved by more than 30% compared to traditional single-closed-loop control, and the pressure control accuracy error is controlled within ±0.1 MPa. This effectively solves the technical defects of traditional control, such as lag and large overshoot, ensuring dynamic self-balancing of casing head pressure and meeting the stringent requirements for pressure control accuracy in ultra-high pressure oil and gas extraction. This invention employs a dual-modal actuator. The main channel high-speed servo hydraulic valve enables rapid response to sudden pressure changes, while the secondary channel precision electrically controlled needle valve achieves high-precision fine-tuning during steady-state operation. The system automatically switches execution modes based on the rate of pressure change. For the first time, this invention integrates high-power hydraulic response with high-precision electric adjustment capabilities in a casing head system, achieving fast-precision coordinated control and filling a gap in the industry. A fault redundancy mechanism is also added to ensure the reliability of the actuator and avoid well control risks caused by execution failures. Based on a real-time temperature-pressure-flow rate three-parameter coupling model, this invention constructs an annular protection fluid injection volume prediction function. The system automatically adjusts the injection pump frequency and flow rate according to this algorithm, achieving second-level dynamic compensation. This avoids the overpressure or underpressure risks caused by traditional timed injection. By incorporating thermodynamic parameters into the injection control logic, it achieves intelligent compensation by controlling pressure with temperature and adjusting fluid with pressure, significantly improving wellbore barrier integrity, reducing the risk of casing damage and oil / gas leakage, and extending wellbore service life. This invention deploys an adaptive learning engine on a cloud server to collect historical operating data from multiple oil well casing heads. It uses machine learning algorithms to train a prediction model and an optimal control strategy library, enabling data interaction and parameter updates between the local control unit and the cloud server. This allows the system to adapt to dynamic changes in different well conditions. As operating data accumulates, prediction accuracy and control effectiveness continuously improve, achieving a "the more you use it, the smarter it becomes" technical effect. At the same time, it reduces network latency through edge computing nodes, solving the problems of lagging data fusion and low intelligence level in existing systems, and reducing manual operation and maintenance costs.

[0063] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A casing head pressure self-balancing and annular pressure management control system, comprising a real-time sensing module, a three-dimensional pressure decoupling module, a dual closed-loop adaptive control module, a dual-modal execution module, an annular protection fluid intelligent injection module, a cloud adaptive learning module, and a control center, characterized in that: The real-time sensing module constructs a three-channel real-time sensing network for casing pressure, annular pressure, and formation pressure to collect pressure, temperature, and flow parameters related to the casing head and wellbore. The three-dimensional pressure decoupling module constructs a three-dimensional pressure decoupling model based on the collected pressure parameters and separates coupled interference signals using an adaptive Kalman filter algorithm, achieving independent identification and dynamic weight allocation of each pressure source. The dual-closed-loop adaptive control module employs a dual-layer closed-loop logic of inner-loop adjustment and outer-loop calibration to achieve dynamic self-balancing of the casing head pressure. The dual-modal execution module executes pressure regulation actions, achieving rapid pressure response and high-precision pressure maintenance. The intelligent injection module for annular protective fluid is connected to the real-time sensing module and the control center, and realizes dynamic injection of annular protective fluid based on a three-parameter coupling model of temperature-pressure-flow rate; the cloud adaptive learning module is connected to the control center, and realizes dynamic updating of control parameters by training a prediction model and control strategy library through machine learning algorithms; the control center coordinates the collaborative work of each module and automatically switches the control mode according to the real-time pressure gradient difference.

2. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The construction process of the three-dimensional pressure decoupling model is as follows: using casing pressure P1, annular pressure P2, and formation pressure P3 as input variables, a three-dimensional pressure coupling equation is constructed. An adaptive Kalman filter algorithm is used to separate the coupling signals, obtaining the independent signals of each pressure source, thus achieving dynamic weight allocation. The state equation of the adaptive Kalman filter algorithm is: Xk = AXk-1 + BUk + Wk, The observation equation is: Zk=HXk+Vk, Where Xk is the state vector at time k, containing the true values ​​of P1, P2, and P3; A is the state transition matrix, a 3×3 identity matrix; B is the control matrix, a 3×1 zero matrix; Uk is the control input at time k, with a value of 0; Wk is the process noise, following a Gaussian distribution with a mean of 0 and a variance of Q, where Q is a 3×3 diagonal matrix with diagonal elements of 0.01, 0.01, and 0.02; Zk is the observation vector at time k, i.e., the pressure value collected by the real-time sensing module; H is the observation matrix, a 3×3 identity matrix; Vk is the observation noise, following a Gaussian distribution with a mean of 0 and a variance of R, where R is a 3×3 diagonal matrix with diagonal elements of 0.005, 0.005, and 0.

01.

3. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The control logic of the dual closed-loop adaptive control module is as follows: the inner loop uses the deviation ΔP between the real-time pressure of the bushing head and the preset pressure threshold as the control target, and adjusts the opening of the balancing valve group and the power of the compensation pump through the pressure self-balancing execution module to achieve rapid pressure response; the outer loop aims at the stability of pressure regulation, collects the pressure fluctuation data after the inner loop adjustment in real time, and dynamically corrects the inner loop control parameters through an adaptive fuzzy PID algorithm. The inner loop control parameters include the proportional coefficient Kp, integral time Ti, and derivative time Td; the parameter correction formula of the adaptive fuzzy PID algorithm is: ΔKp=Kp0+Kp1×E+Kp2×EC, ΔTi = Ti0 + Ti1 × E + Ti2 × EC ΔTd = Td0 + Td1 × E + Td2 × EC; Wherein, ΔKp, ΔTi, and ΔTd are the correction amounts for the proportional coefficient, integral time, and derivative time, respectively; Kp0, Ti0, and Td0 are the initial control parameters, with values ​​of 5.0, 1.2s, and 0.3s, respectively; Kp1, Kp2, Ti1, Ti2, Td1, and Td2 are the correction coefficients, with values ​​of 0.8, 0.4, 0.1, 0.05, 0.08, and 0.04, respectively; E is the fuzzy quantized value of the pressure deviation ΔP, and EC is the fuzzy quantized value of the pressure deviation change rate.

4. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The dual-mode execution module includes a main channel and a secondary channel. The main channel is a high-speed servo hydraulic valve with a response time ≤50ms, used for rapid pressure relief or replenishment during sudden pressure changes. The secondary channel is a precision electrically controlled needle valve driven by a stepper motor, with a pressure regulation resolution of 0.01MPa, used for high-precision pressure maintenance in steady state. The control center automatically switches the execution mode according to the pressure change rate v: when v > 0.3MPa / s, the main channel high-speed servo hydraulic valve is activated for rapid adjustment; when v < 0.05MPa / s, it switches to the secondary channel precision electrically controlled needle valve for fine adjustment; when 0.05MPa / s ≤ v ≤ 0.3MPa / s, the two channels work together. The dual-mode execution module also has a fault redundancy module. When either channel fails, the other channel automatically switches to the main control channel to ensure continuous pressure regulation.

5. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The intelligent injection module for annular protection fluid is based on a real-time temperature-pressure-flow rate three-parameter coupling model, and constructs an annular protection fluid injection volume prediction function. The prediction function is as follows: Q = K × (P × T) / F; Where Q is the annular protective fluid injection volume, in L / min; K is the injection coefficient, ranging from 0.02 to 0.05, dynamically adjusted according to well conditions; P is the real-time annular pressure, in MPa; T is the real-time annular temperature, in °C; and F is the protective fluid viscosity coefficient, in mPa·s, preset according to the protective fluid type. The system automatically adjusts the injection pump frequency and flow rate based on this prediction function to achieve second-level dynamic compensation, with an injection response time ≤1s.

6. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The cloud adaptive learning module includes a cloud server and a local data interaction unit. The cloud server deploys an adaptive learning engine to collect historical operating data from multiple well casing heads. This historical operating data includes annular pressure variation curves, temperature variation curves, control action records, and corresponding production condition parameters. The adaptive learning engine uses a random forest algorithm to train a dynamic annular pressure behavior prediction model. The prediction formula for the random forest algorithm is: Y = (1 / N) × ΣYi; Where Y is the predicted annular pressure value in MPa; N is the number of decision trees, ranging from 100 to 200; Yi is the predicted value of the i-th decision tree in MPa; and the gradient boosting decision tree algorithm is used to train the optimal control strategy library to achieve adaptive matching of control parameters under different well conditions.

7. The casing head pressure self-balancing and annular pressure management control system according to claim 6, characterized in that: The local data interaction unit establishes a connection with the cloud server via industrial Ethernet or 5G wireless communication technology, and periodically downloads updated prediction model parameters and control strategy parameters from the cloud server. The update cycle is 1 to 24 hours, dynamically adjusted according to the complexity of the well conditions. At the same time, the local data interaction unit uploads the actual effect data of each pressure regulation to the cloud server. The actual effect data includes the pressure value before and after regulation, the regulation time, and the overshoot, which serves as training samples for continuous model optimization. The cloud server is also equipped with a data encryption module to ensure the security of data transmission and storage.

8. The casing head pressure self-balancing and annular pressure management control system according to claim 3, characterized in that: The control center has two control modes: self-balancing dominant and annular compensation dominant. The control center automatically switches modes based on the real-time pressure deviation value ΔP: when ΔP < 0.02 MPa / m, it switches to self-balancing dominant mode, and the pressure regulation is dominated by the dual closed-loop adaptive control module. When ΔP≥0.02MPa / m, the system switches to the annular compensation-dominated mode, where the intelligent injection module for annular protective fluid dominates pressure regulation, and the dual closed-loop adaptive control module assists in regulation. The control center is also equipped with an alarm module, which automatically issues an audible and visual alarm and triggers an emergency pressure relief action when the pressure parameter exceeds the preset safety threshold ±0.2MPa.

9. The casing head pressure self-balancing and annular pressure management control system according to claim 1, characterized in that: The real-time sensing module includes a casing pressure sensor, annular pressure sensor, formation pressure sensor, temperature sensor, and flow sensor. Each sensor has a sampling frequency of 10~50Hz, a pressure measurement accuracy of ±0.01MPa, a temperature measurement accuracy of ±0.1℃, and a flow measurement accuracy of ±0.1L / min. The sensors are designed to be corrosion-resistant and resistant to high pressure, with an operating pressure range of 0~100MPa and an operating temperature range of -40℃~150℃, making them suitable for extreme downhole conditions in oilfields.

10. A casing head pressure self-balancing and annular pressure management and control system according to any one of claims 1-9, characterized in that: The system also includes a data storage module and a remote monitoring module. The data storage module stores real-time sensor data, control parameters, and fault records. The remote monitoring module allows administrators to view the system's operating status and parameter curves in real time through terminal devices, and provides the function of remotely issuing control commands to achieve remote operation and maintenance.

Images