Well completion flow redundancy control method and device

Abstract

The embodiment of the invention provides a well completion flow redundancy control method and device, and the method comprises the steps: collecting underground environment data in real time, inputting the underground environment data into a Gaussian process model to predict the change trend of the underground environment, generating a flow control instruction based on the change trend of the underground environment, and carrying out the rolling optimization of the flow control instruction. The optimized flow control instruction is executed according to the dual-flow control valve deployed underground, the state of the flow control valve can be automatically adjusted, accurate control and redundancy management of the flow can be achieved so as to adapt to underground complex and changeable production conditions, and the safety and efficiency of oil and gas production are improved.

Description

Technical Field

[0001] The invention relates to the technical field of petroleum engineering, and in particular to a method and device for redundant control of well completion flow. Background Art

[0002] With the continuous development of the petroleum industry, more stringent technical and efficiency requirements have been put forward for the development and management of oil and gas wells. Oil and gas well completion operations are a crucial link in the oilfield production process, and flow control is directly related to production safety and optimization of oil and gas production.

[0003] Traditional flow control methods mainly rely on the experience of on-site operators. This manual adjustment method has many challenges. First, manual adjustment based on experience makes the response speed of flow control relatively slow, and it is difficult to quickly adapt to the variability of the downhole production environment. Secondly, due to the complexity of the geological conditions of oil and gas wells, operators may find it difficult to accurately judge the actual situation downhole, resulting in inaccurate flow control. In addition, the manually adjusted flow control method also has the problems of complex operation and long time consumption, which affects the overall production efficiency of oil and gas wells.

[0004] Therefore, there is an urgent need for an innovative technology that can realize intelligent and automated flow control to meet the needs of the modern petroleum industry for efficient and safe production. Summary of the invention

[0005] In view of the above problems, a completion flow redundancy control method and device are proposed to overcome the above problems or at least partially solve the above problems, including:

[0006] Real-time collection of underground environmental data;

[0007] Input downhole environmental data into the Gaussian process model to predict the downhole environmental change trend;

[0008] Generate flow control instructions based on the changing trend of the underground environment, and perform rolling optimization on the flow control instructions;

[0009] The optimized flow control instructions are executed according to the dual flow control valves deployed underground.

[0010] Preferably, real-time collection of downhole environmental data includes:

[0011] Using sensors deployed underground, the pressure data, fluid data and temperature data of the underground are collected in real time; among which, the fluid data includes fluid density, fluid velocity and flow rate.

[0012] Preferably, before inputting the downhole environmental data into the Gaussian process model, the method further includes: constructing a tubing pressure drop calculation model, an annular space pressure drop calculation model and a temperature calculation model according to the downhole environmental data;

[0013] Among them, the calculation model of oil pipe pressure drop is:

[0014]

[0015] The calculation model of annular space pressure drop is:

[0016]

[0017] In the formula, subscripts a and t represent the annulus and tubing respectively, subscript m represents the properties of the mixed fluid, f represents the friction coefficient of the pipe wall, ρ represents the fluid density, and v represents the fluid flow rate;

[0018] The temperature calculation model is:

[0019]

[0020] Where T is temperature, ρ is density, c p is the specific heat capacity, k is the thermal conductivity, z is the depth coordinate, and Q is the heat source.

[0021] Preferably, the Gaussian process is fully specified by a mean function m() and a covariance function k(,), defined as:

[0022]

[0023] Preferably, rolling optimization for the flow control instruction includes:

[0024] The relationship between the prediction error and the state and input effect is established, the Gaussian process model is compensated according to the prediction error, and the optimized Gaussian process model is used to re-predict the downhole environmental change trend and generate new flow control instructions.

[0025] Preferably, the relationship between the prediction error and the state and input effect is as follows:

[0026]

[0027] In the formula, d is the prediction error, x is the intelligent completion temperature, pressure, and flow rate, and u is the flow valve opening; is the unmodeled part of the well completion, w is the external interference, and f is the modeled part, namely the tubing pressure drop calculation model, the annular space pressure drop calculation model, and the temperature calculation model.

[0028] Preferably, the Gaussian process model selects the square exponential kernel function:

[0029]

[0030] Preferably, the method further comprises optimizing hyperparameters by maximizing a marginal likelihood function.

[0031] Preferably, executing the optimized flow control instruction according to the dual flow control valve deployed underground includes:

[0032] The first flow control valve is controlled to execute the optimized flow control instruction. When the first flow control valve fails, the second flow control valve is controlled to execute the optimized flow control instruction.

[0033] A well completion flow redundancy control device, the device comprising:

[0034] Acquisition module, used to collect downhole environmental data in real time;

[0035] A prediction module, used for inputting downhole environmental data into a Gaussian process model to predict the downhole environmental change trend;

[0036] An instruction generation module is used to generate flow control instructions based on the change trend of the underground environment and perform rolling optimization on the flow control instructions;

[0037] The execution module is used to execute the optimized flow control instruction according to the dual flow control valve deployed underground.

[0038] The embodiment of the present invention collects downhole environmental data in real time, inputs the downhole environmental data into a Gaussian process model to predict the downhole environmental change trend, generates flow control instructions based on the downhole environmental change trend, and performs rolling optimization on the flow control instructions. The optimized flow control instructions are executed according to the dual flow control valves deployed downhole, and the flow control valve state is automatically adjusted to achieve precise control and redundant management of the flow, so as to adapt to the complex and changeable production conditions downhole and improve the safety and efficiency of oil and gas production. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0040] Figure 1 It is a flowchart of the steps of a well completion flow redundancy control method provided by the present invention;

[0041] Figure 2 It is a flow chart for solving a pressure and temperature model provided by the present invention;

[0042] Figure 3 It is a flow chart of updating the pressure and temperature model parameters provided by the present invention;

[0043] Figure 4It is a schematic diagram of the overall framework of the well completion flow redundancy control provided by the present invention. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Reference Figure 1 , shows a flowchart of a method for controlling well completion flow redundancy provided by an embodiment of the present invention, which may specifically include the following steps:

[0046] Step 101, collecting downhole environmental data in real time.

[0047] In the embodiments of the present invention, various types of sensors, such as pressure sensors, flow sensors, and temperature sensors, can be deployed underground, and then the sensors deployed underground can be used to collect underground environmental data in real time. Specifically, underground pressure data, fluid data, and temperature data can be collected; wherein the fluid data may include fluid density, fluid velocity, and flow rate, etc.

[0048] In order to obtain accurate temperature and pressure distribution of well completion, the embodiment of the present invention establishes a well completion coupled pressure and temperature model:

[0049] Specifically, based on the pressure and temperature change mechanism of multiphase flow in the wellbore, the mass conservation equation, momentum conservation equation and energy conservation equation can be used to establish the tubing pressure drop calculation model, the annular space pressure drop calculation model and the temperature calculation model for wellbore flow.

[0050] Firstly, the momentum conservation equation can be established based on the fact that the force on the fluid in the volume element is equal to the momentum change of the fluid in the volume element, and on this basis, the pressure drop calculation model of the wellbore can be obtained.

[0051] The calculation model of the oil pipe pressure drop is:

[0052]

[0053] The pressure drop calculation model of the annular space is:

[0054]

[0055] Wherein, subscripts a and t represent the annulus and tubing, respectively; subscript m represents the properties of the mixed fluid; f represents the friction coefficient of the pipe wall; represents the fluid density; and v represents the fluid flow rate.

[0056] Secondly, the temperature model of the wellbore is derived based on the mass conservation equation and energy conservation equation of the volume element. The temperature change mainly occurs along the depth direction of the wellbore, and the one-dimensional heat conduction equation is used to describe this situation:

[0057]

[0058] Where: T is temperature, ρ is density, c p is the specific heat capacity, k is the thermal conductivity, z is the depth coordinate, and Q is the heat source, such as Figure 2 , 3 , showing the pressure and temperature model solution and parameter update flow chart.

[0059] Step 102, inputting the downhole environmental data into the Gaussian process model to predict the downhole environmental change trend.

[0060] In the embodiment of the present invention, the pressure and temperature data provided by the tubing pressure drop calculation model, the annular space pressure drop calculation model and the temperature calculation model are important inputs of the Gaussian process model. The Gaussian process model can predict the changing trend of the downhole environment by learning these data.

[0061] Specifically, similar to the multivariate Gaussian distribution represented by the mean vector μ and the covariance matrix ∑, the Gaussian process is completely specified by the mean function m() and the covariance function k(,), defined as:

[0062]

[0063] Among them, m(z) is the mean function, which represents the expected value of the Gaussian process, and k(z,z′) is the covariance function, which represents the correlation between two points z and z′.

[0064] Furthermore, the Gaussian process model can predict the changing trend of the downhole environment by modeling the pressure and temperature data. The prediction error d(k) is defined as:

[0065]

[0066] Where d is the prediction error, x is the completion status (temperature, pressure, flow), and u is the flow valve opening. is the unmodeled part of the intelligent completion, w is the external disturbance, and f is the modeled part, that is, the completion temperature and pressure model. It can be understood that this formula can also reflect the relationship between the prediction error and the state and input effect.

[0067] In addition, the embodiment of the present invention also measures Defined as:

[0068]

[0069] Dataset Defined as:

[0070]

[0071] Where, z(k): = [x(k), u(k)] T is the combination of state and input, d(k) is the prediction error, and N is the number of data points.

[0072] Therefore, it is assumed that the prediction error vector d satisfies the following Gaussian distribution:

[0073]

[0074] To avoid expensive posterior computation, we assume, without loss of generality, that the prior GP mean m(z) = 0, and inference is performed only through the covariance function. Empirically, setting the prior to 0 is often accomplished by subtracting the (prior) mean from all observations.

[0075] In addition, kernel functions play a vital role in Gaussian process models. Kernel functions are used to measure the similarity or correlation between two data points, thereby defining the covariance structure of the Gaussian process. The selection of a suitable kernel is based on assumptions such as smoothness and possible patterns in the data. A reasonable assumption is usually that the correlation between two points decays with the distance between the points. This means that points that are closer behave more similarly than points that are farther away. In the Gaussian process model in the embodiment of the present invention, a square exponential kernel function is selected, which is defined as:

[0076]

[0077] Among them, z i and z j is the input point, is the signal variance, and the hyperparameter λ is the length-scale. Specifically, Control function volatility: Larger Indicates that the function has a larger range of variation, and a smaller Indicates that the function has a smaller range of variation. λ controls the correlation between points: a larger λ means that the correlation between points decays more slowly, and a smaller λ means that the correlation decays faster.

[0078] The two hyperparameters λ (called the length-scale) and The signal variance can be changed to increase or decrease the prior correlation between points, thereby increasing or decreasing the variability of the resulting function.

[0079] It can be understood that through the kernel function, the Gaussian process model can generate the predicted mean and variance of new data points, thereby realizing the prediction of future states.

[0080] In the embodiment of the present invention, the hyperparameters determine the flexibility of the kernel function, which in turn affects the prediction performance of the model. In order to train the Gaussian process regression model, it is necessary to estimate the hyperparameters involved in the kernel function k.

[0081] It is understandable that the goal of hyperparameter optimization is to find a set of optimal hyperparameters η = [λ,σ f ], so that the prediction error of the Gaussian process model is minimized. In an embodiment of the present invention, the hyperparameters can be estimated by maximizing the marginal likelihood function.

[0082] Specifically, the marginal likelihood function can be expressed as:

[0083]

[0084] where d is the prediction error vector, K(Z, Z) is the covariance matrix between the training data points, and the kernel function k(z i ,z j ) is calculated, is the variance of the observation noise and I is the identity matrix.

[0085] The logarithmic form of the marginal likelihood function is:

[0086]

[0087] Where N is the number of data points, is the determinant of the covariance matrix.

[0088] By maximizing the marginal likelihood function, the optimal hyperparameter η* can be found:

[0089] η*=argmax η logp(d|Z,η)

[0090] In the optimization process, the noise variance is also unknown and needs to be estimated in the actual process. The mismatch model can be estimated using the least squares method, and then the sampling variance is used as the estimate of the noise variance.

[0091] Specifically, using the sampling variance as an estimated value of the noise variance may include the following steps:

[0092] 1. Calculate the variance of the observed data:

[0093]

[0094] in, is the mean of the observed data.

[0095] 2. Estimate the noise variance:

[0096] σ 2=Var(y)

[0097] Through noise estimation, the Gaussian process model can better handle the noise in the observed data, thereby improving the predictive performance and robustness of the model.

[0098] In an embodiment of the present invention, the training data And the observed value d and function value of the new data point z Obey a joint (multivariate) normal distribution:

[0099]

[0100] In the formula, is the covariance matrix between all observation points so far, K(z,z) is the covariance matrix between the newly introduced points, is the covariance matrix between the new input points and the observed points, is the covariance matrix between the observed points and the new input points. In addition, I is the identity matrix, is the assumed noise level of the observations.

[0101] In the embodiment of the present invention, the posterior is also a GP with a mean function and a covariance function, and its mean μ * and covariance function Σ * Defined as:

[0102]

[0103] In the embodiment of the present invention, it is also related to how to propagate the uncertainty of the input data to the prediction result of the model. Solving the uncertainty propagation problem is crucial to improving the robustness and reliability of the model.

[0104] When the Gaussian process model is paired with the nominal system model f, the covariance between the state, input, and Gaussian process model is considered, that is, ∑ x and∑ u , which is very important for accurate propagation of uncertainty. Therefore, the nominal system dynamics can be linearized around the mean:

[0105]

[0106] Furthermore, the embodiment of the present invention can calculate the covariance matrix of the prediction result through the covariance propagation formula, thereby realizing uncertainty propagation, specifically:

[0107] Assume that the covariance matrices of the state variables and control inputs are Σ x and Σ u , then the covariance matrix of the prediction result is Σ f It can be calculated by the following formula:

[0108]

[0109] Among them, ▽f(μ x ,μ u ) is the model at the mean point (μ x ,μ u ) is the gradient at .

[0110] In an embodiment of the present invention, by considering the uncertainty of input data, the model can generate more robust prediction results, thereby maintaining stability in the face of uncertainty, and the system can better understand the uncertainty of the prediction results and generate more reliable decisions.

[0111] In addition, it can be understood that the update equation is the core part of the combination of model predictive control and Gaussian process model. Specifically, the update equation is used to describe the update process of the mean and covariance matrix of the system state at each time step. The goal of the update equation is to predict the state mean and covariance matrix of the next time step based on the current state and control input.

[0112] In the embodiment of the present invention, the update equations of the state mean and variance are:

[0113]

[0114] Among them, ∑ k is the joint covariance matrix of the current state and the control input:

[0115]

[0116] By updating the equations, the system can predict the state mean and covariance moment of the next time step based on the current state and control input. It describes how to propagate the uncertainty of the current state and control input to the next time step, thereby generating a more robust prediction result.

[0117] Step 103, generating a flow control instruction based on the downhole environment change trend, and performing rolling optimization on the flow control instruction.

[0118] In the embodiment of the present invention, a flow control instruction suitable for the current state can be generated based on the prediction results of the Gaussian process model. Since the underground environment changes in real time, it is necessary to optimize the flow control instruction, or evaluate the effects of different flow control instructions by calculating the value function, and select the optimal control input. Specifically:

[0119] The value function is defined as:

[0120]

[0121] In the formula, is the terminal state mean, The terminal state variance, N is the prediction time domain, is the mean value of the i-th state quantity, The mean of the i-th input quantity, The variance of the i-th state variable.

[0122] In another embodiment of the present invention, the prediction results of multiple models may be dynamically selected or integrated to generate the final control instruction according to the current downhole conditions and the performance of the model.

[0123] The embodiment of the present invention can also compensate for the Gaussian process model according to the prediction error, and use the optimized Gaussian process model to re-predict the change trend of the underground environment and generate new flow control instructions.

[0124] Step 104, executing the optimized flow control instruction according to the dual flow control valve deployed underground.

[0125] In an embodiment of the present invention, a dual flow control valve can be deployed underground, and automatically adjusted by an intelligent control system according to the decision instructions of the deep learning model. If the main actuator fails, the system will automatically switch to the backup actuator.

[0126] Specifically, controlling the first flow control valve to execute the optimized flow control instruction, and in the event of a failure of the first flow control valve, controlling the second flow control valve to execute the optimized flow control instruction can ensure that the execution of the flow control instruction is not affected, thereby further ensuring the production efficiency of the oil and gas well.

[0127] In addition, the embodiments of the present invention can also continuously monitor sensor data, control valve status and communication networks. Once an abnormality or failure is detected, the backup system is immediately started and the failed component is automatically restored, achieving redundant control of the completion flow and providing multiple guarantees for the workflow. Figure 4 , showing the overall framework diagram of completion flow redundancy control.

[0128] The embodiment of the present invention provides a completion flow redundancy control method and device, which collects downhole environmental data in real time, inputs the downhole environmental data into a Gaussian process model to predict the downhole environmental change trend, generates flow control instructions based on the downhole environmental change trend, and performs rolling optimization on the flow control instructions. The optimized flow control instructions are executed according to the dual flow control valves deployed downhole, and the flow control valve state can be automatically adjusted to achieve precise control and redundant management of the flow, so as to adapt to the complex and changeable production conditions downhole and improve the safety and efficiency of oil and gas production.

[0129] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0130] The present invention also provides an embodiment of a well completion flow redundancy control device, which may specifically include the following modules:

[0131] Acquisition module, used to collect downhole environmental data in real time;

[0132] A prediction module, used for inputting downhole environmental data into a Gaussian process model to predict the downhole environmental change trend;

[0133] An instruction generation module is used to generate flow control instructions based on the change trend of the underground environment and perform rolling optimization on the flow control instructions;

[0134] The execution module is used to execute the optimized flow control instruction according to the dual flow control valve deployed underground.

[0135] Preferably, real-time collection of downhole environmental data includes:

[0136] Using sensors deployed underground, the pressure data, fluid data and temperature data of the underground are collected in real time; among which, the fluid data includes fluid density, fluid velocity and flow rate.

[0137] Preferably, before inputting the downhole environmental data into the Gaussian process model, the method further includes: constructing a tubing pressure drop calculation model, an annular space pressure drop calculation model and a temperature calculation model according to the downhole environmental data;

[0138] Among them, the calculation model of oil pipe pressure drop is:

[0139]

[0140] The calculation model of annular space pressure drop is:

[0141]

[0142] In the formula, subscripts a and t represent the annulus and tubing respectively, subscript m represents the properties of the mixed fluid, f represents the friction coefficient of the pipe wall, ρ represents the fluid density, and v represents the fluid flow rate;

[0143] The temperature calculation model is:

[0144]

[0145] Where T is temperature, ρ is density, c p is the specific heat capacity, k is the thermal conductivity, z is the depth coordinate, and Q is the heat source.

[0146] Preferably, the Gaussian process is fully specified by a mean function m() and a covariance function k(,), defined as:

[0147]

[0148] Preferably, rolling optimization for the flow control instruction includes:

[0149] The relationship between the prediction error and the state and input effect is established, the Gaussian process model is compensated according to the prediction error, and the optimized Gaussian process model is used to re-predict the downhole environmental change trend and generate new flow control instructions.

[0150] Preferably, the relationship between the prediction error and the state and input effect is as follows:

[0151]

[0152] In the formula, d is the prediction error, x is the intelligent completion temperature, pressure, and flow rate, and u is the flow valve opening; is the modeling part for well completion, w is the external interference, and f is the modeled part, namely the tubing pressure drop calculation model, the annular space pressure drop calculation model, and the temperature calculation model.

[0153] Preferably, the Gaussian process model selects the square exponential kernel function:

[0154]

[0155] Preferably, the method further comprises optimizing hyperparameters by maximizing a marginal likelihood function.

[0156] Preferably, executing the optimized flow control instruction according to the dual flow control valve deployed underground includes:

[0157] The first flow control valve is controlled to execute the optimized flow control instruction. When the first flow control valve fails, the second flow control valve is controlled to execute the optimized flow control instruction.

[0158] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0159] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0160] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0161] The above is a detailed introduction to a completion flow redundancy control method and device provided. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technicians in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for redundant control of well completion flow, characterized in that: The method comprises: Real-time collection of underground environmental data; Inputting the downhole environmental data into a Gaussian process model to predict the downhole environmental change trend; Generate a flow control instruction based on the downhole environment change trend, and perform rolling optimization on the flow control instruction; The optimized flow control instructions are executed according to the dual flow control valves deployed underground.

2. The method according to claim 1, characterized in that The real-time acquisition of downhole environmental data includes: Using sensors deployed underground, the pressure data, fluid data and temperature data of the underground are collected in real time; wherein the fluid data includes fluid density, fluid velocity and flow rate.

3. The method according to claim 2, characterized in that Before inputting the downhole environmental data into the Gaussian process model, the method further includes: constructing a tubing pressure drop calculation model, an annular space pressure drop calculation model and a temperature calculation model according to the downhole environmental data; Among them, the calculation model of oil pipe pressure drop is: The calculation model of annular space pressure drop is: In the formula, subscripts a and t represent the annulus and tubing respectively, subscript m represents the properties of the mixed fluid, f represents the friction coefficient of the pipe wall, ρ represents the fluid density, and v represents the fluid flow rate; The temperature calculation model is: Where T is temperature, ρ is density, c p is the specific heat capacity, k is the thermal conductivity, z is the depth coordinate, and Q is the heat source.

4. The method according to claim 3, characterized in that The Gaussian process is fully specified by the mean function m() and the covariance function k(,), defined as:

5. The method according to claim 4, characterized in that The rolling optimization for the flow control instruction comprises: The relationship between the prediction error and the state and input effect is established, the Gaussian process model is compensated according to the prediction error, and the optimized Gaussian process model is used to re-predict the downhole environment change trend and generate new flow control instructions.

6. The method according to claim 5, characterized in that The relationship between the prediction error and the state and input effect is as follows: In the formula, d is the prediction error, x is the intelligent completion temperature, pressure, and flow rate, and u is the flow valve opening; is the unmodeled part of the well completion, w is the external interference, and f is the modeled part, namely the tubing pressure drop calculation model, the annular space pressure drop calculation model, and the temperature calculation model.

7. The method according to claim 6, characterized in that The Gaussian process model selects the square exponential kernel function:

8. The method according to claim 7, characterized in that The method also includes optimizing the hyperparameters by maximizing a marginal likelihood function.

9. The method according to claim 1, characterized in that: The method of executing the optimized flow control instruction according to the dual flow control valve deployed underground includes: The first flow control valve is controlled to execute the optimized flow control instruction. When the first flow control valve fails, the second flow control valve is controlled to execute the optimized flow control instruction.

10. A well completion flow redundancy control device, characterized in that: The device comprises: Acquisition module, used to collect downhole environmental data in real time; A prediction module, used for inputting the downhole environment data into a Gaussian process model to predict the downhole environment change trend; An instruction generation module, used for generating a flow control instruction based on the downhole environment change trend, and performing rolling optimization on the flow control instruction; The execution module is used to execute the optimized flow control instruction according to the dual flow control valve deployed underground.

Images