Fractured well single well production regulation and control method and system based on real-time diagnosis

Abstract

The invention relates to a fractured well single well production regulation and control method and system based on real-time diagnosis, and aims to overcome the defects of strategy staticizing, diagnosis feedback lagging, poor field data quality and the like in the prior art. The method comprises the following steps: acquiring flowing bottomhole pressure data and yield data of a target production well in real time; after the data are preprocessed, the effective fracture surface area representing the fracture conductivity is obtained through inversion calculation; and according to the real-time change trend of the effective crack surface area, a regulation and control instruction for adjusting the opening degree of the oil nozzle is automatically generated. According to the invention, through virtualized data acquisition and high-fidelity preprocessing, continuous and reliable supply of key data at low cost is realized; a global transient analysis model is engineered into a real-time diagnosis core, and online accurate sensing of the crack health state is achieved; and finally, a closed loop of perception-diagnosis-decision-control is formed, self-adaptive production regulation and control of a single well are achieved, fracture conductivity attenuation is effectively delayed, and the final recovery efficiency is improved.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of unconventional oil and gas field development and intelligent control, specifically relating to a single-well production control method and system for fracturing wells such as shale oil and gas wells. Background Technology

[0002] Unconventional oil and gas resources, such as shale oil and gas, have become an important part of the global energy supply. These reservoirs typically have extremely low matrix permeability, therefore their development is highly dependent on horizontal well drilling technology and large-scale hydraulic fracturing technology. Creating complex artificial fracture networks in tight reservoirs through hydraulic fracturing is a prerequisite for the commercial exploitation of shale oil and gas.

[0003] However, shale reservoirs are typically highly stress-sensitive. During production, as formation fluids are extracted, pore pressure decreases, leading to a significant increase in effective stress acting on the rock skeleton and fracture surfaces. This increase in effective stress can cause proppant embedding and breakage, as well as the closure of unsupported fractures, resulting in irreversible damage to fracture conductivity (or effective fracture surface area). Therefore, developing a scientifically sound production system (i.e., precisely controlling nozzle opening) is crucial for delaying the decline in fracture conductivity and improving the ultimate recovery rate (EUR) of a single well.

[0004] Current shale oil and gas well production control technologies have the following limitations:

[0005] First, the control strategies are static and lack individual well-specificity. Existing "pressure-controlled production" systems are typically based on static geological parameters of the block or limited numerical simulation results, setting a uniform production pressure differential window (e.g., stipulating that the daily wellhead pressure drop should not exceed a specific value). This "one-size-fits-all" management model ignores the strong geological heterogeneity and the differences in fracturing effects among individual wells. For wells with good geological conditions and stable fracture support, excessive pressure control can limit production release; while for wells with extremely high stress sensitivity, a uniform pressure drop rate may still lead to premature fracture closure. Furthermore, this model cannot adapt to the dynamic changes in production systems required by individual wells at different life stages (such as the early linear flow stage and the later boundary control flow stage).

[0006] Secondly, feedback is delayed and relies on manual analysis. Currently, the evaluation of fracture conductivity or effective fracture area mainly depends on production transient analysis (RTA) or well test analysis (PTA). These traditional methods typically require exporting production data accumulated over several weeks or months for offline analysis by professional engineers. Faced with complex operating conditions such as frequent well opening and closing and nozzle adjustments in the field, traditional analysis charts (such as log-log plots or square root time plots) often become ineffective due to severe data scattering, making it difficult to draw accurate conclusions. Although the Global Transient Analysis (UTA) theory proposed in recent years has shown advantages in processing variable flow data, it is still mainly at the stage of theoretical modeling or post-analysis, lacking supporting automated denoising, segmentation, and feature recognition algorithms, and cannot meet the closed-loop control requirements of "real-time diagnosis and real-time feedback" in industrial fields.

[0007] Furthermore, the low quality of field multiphase flow data affects the accuracy of inversion. Accurate fracture parameter inversion highly depends on high-quality bottom-hole flowing pressure (BHP) and three-phase production data of oil, gas, and water. However, due to cost considerations, most shale oil and gas wells do not have permanent downhole pressure gauges installed, and expensive multiphase flow meters (MPFM) or test separators are lacking on the surface. Existing production data often contains a large amount of noise, outliers, and data gaps due to intermittent manual metering. Existing control systems generally lack effective data completion (such as virtual metering) and high-fidelity cleaning mechanisms, and direct analysis based on poor-quality data often leads to erroneous decisions.

[0008] In view of this, there is an urgent need for an intelligent method and system that can use low-cost conventional surface monitoring data to invert the health status of downhole fractures in real time and accurately, and automatically adjust the production system accordingly, so as to achieve refined management of each production well and optimization of production capacity throughout its entire life cycle. Summary of the Invention

[0009] In view of this, the purpose of this invention is to provide a single-well closed-loop control method and system that can utilize low-cost conventional surface monitoring data to achieve real-time and accurate inversion of the downhole fracture health status, and automatically and intelligently adjust the production system accordingly, so as to overcome the problems of static control, feedback lag and poor data quality, and ultimately achieve single-well full life cycle production optimization.

[0010] The first aspect of this invention discloses a method for controlling the production of a single-well fractured well based on real-time diagnosis, comprising the following steps: S1, acquiring the single-well production data of the target production well in real time, wherein the single-well production data includes pressure data and production data; S2, preprocessing the single-well production data; S3, obtaining the effective fracture surface area characterizing the fracture conductivity through inversion calculation based on the preprocessed single-well production data; S4, generating and executing a control command for adjusting the nozzle opening of the target production well according to the real-time change trend of the effective fracture surface area.

[0011] This invention abandons the traditional static control logic based on fixed pressure drop rates or simple pressure thresholds, and instead uses the effective fracture surface area—a parameter that directly and essentially reflects the dynamic changes of the underground fracture system—as the core basis for control. By acquiring and processing data in real time to continuously invert the effective fracture surface area and using its changing trend as decision input, a fundamental shift from "open-loop empirical control" to "closed-loop state feedback control" is achieved. This scheme provides a feasible technical path for achieving adaptive dynamic optimization control for single wells.

[0012] In other words, by obtaining the effective fracture surface area that directly reflects the internal state of the fracture through real-time inversion, and using this as the core for decision-making, the traditional "one-size-fits-all" static pressure control mode has been completely changed. The system can dynamically adjust according to the real-time status of each well, explore ways to increase production when the fractures are healthy, and provide timely protection when they are damaged, achieving true "one well, one policy" and significantly improving the production utilization rate and final recovery rate of a single well.

[0013] Further, step S1 includes: S101, acquiring the wellhead pressure data of the target production well in real time; S102, acquiring the total equivalent production as one of the production data, which can be obtained based on real-time measurement, or, based on the wellhead pressure data, calculating the continuous estimated production of the oil, gas, and water phases through a virtual flowmeter model, and then calculating the total equivalent production based on the continuous estimated production of the three phases; S103, acquiring the bottomhole flowing pressure data as one of the pressure data, which can be obtained based on real-time monitoring, or, based on the wellhead pressure data and the total equivalent production through back-calculation and reconstruction. In a more preferred embodiment, in step S101, the wellhead pressure data and wellhead temperature data of the target production well can be acquired in real time, and then in step S102, based on the wellhead pressure data and wellhead temperature data, the continuous estimated production of the oil, gas, and water phases can be calculated through a virtual flowmeter model, and then the total equivalent production can be calculated based on the continuous estimated production of the three phases, and then step S03 can be continued.

[0014] Based on this solution, the bottom hole flowing pressure and total equivalent production required for subsequent diagnosis can no longer be obtained solely through expensive direct monitoring. By introducing a "virtual flowmeter model" and "back-calculation reconstruction" technology, the required key parameters can be indirectly and continuously calculated using the lowest-cost and most readily available wellhead pressure and temperature data, combined with known wellbore physical models. This breaks the technical prejudice that "high-quality data must rely on high-cost hardware." Its beneficial effects are direct and significant: it greatly reduces the economic threshold and hardware dependence for system implementation, making the real-time diagnosis and control solution of this invention feasible for large-scale field deployment, thereby effectively solving the problems of poor data quality and high cost in existing technologies.

[0015] Furthermore, the preprocessing steps for the single-well production data include a data cleaning step based on a two-stage adaptive denoising algorithm, comprising: S201, setting a dynamic threshold based on the local variance and statistical distribution characteristics of the single-well production data to remove gross errors; S202, using an adaptive window smoothing filter combined with a median filter to perform secondary smoothing processing on the error-removed data. That is, the first stage (S201) uses statistical methods to identify and remove non-physical gross errors caused by sensor transient failures, signal interference, etc., which can severely distort data trends; the second stage (S202) uses combined filtering techniques to suppress high-frequency random noise while retaining the true pressure / production transient characteristics at different time scales through an adaptive window mechanism. The median filter has good robustness to impulse noise, while adaptive smoothing avoids the loss of details or insufficient smoothing caused by a fixed window. This significantly improves the signal-to-noise ratio of the original data, providing a clean and reliable input for the subsequent high-precision inversion model, which is a fundamental prerequisite for ensuring the diagnostic accuracy of the entire system.

[0016] Furthermore, in the preprocessing step of single-well production data, after the data cleaning step based on the two-stage adaptive denoising algorithm, the step of automatically identifying and segmenting the denoised data into flow states is also included, which includes: S203, calculating the time-varying derivative and normalized rate of change of the single-well production data; S204, when the time-varying derivative and normalized rate of change exceed the adaptive threshold, it is determined as a change point of the production flow state; S205, based on the change point, the continuous time series is automatically divided into several independent flow analysis segments, and invalid data segments with a time length less than a preset value are removed.

[0017] This addresses the problem of diagnostic model failure caused by the mixing of multiple flow states (such as linear flow, boundary flow, and well shut-in recovery) in continuous production data streams. It utilizes the rate of change (derivative) of data as a sensitive indicator of flow regime changes. When operations such as well opening, well shut-in, or significant production adjustments occur, the instantaneous rate of change of pressure or production changes abruptly. By setting an adaptive threshold, these "change points" can be automatically captured. Subsequently, the data is segmented based on these change points, and excessively short or unstable transition segments are eliminated, thus providing a single, stable flow analysis period for subsequent diagnostic models (such as UTA). This enables intelligent and automated segmentation processing of complex production histories, allowing advanced diagnostic models to be applied to continuous real-time data streams containing multiple operating conditions. This is a key step in achieving "real-time diagnosis" and directly solves the feedback lag problem described in the background technology.

[0018] The two-stage adaptive denoising and automatic flow segmentation algorithm in this application effectively overcomes the interference of high noise and complex working conditions in the field data on the high-level model, enabling the whole system to operate stably and reliably in a real industrial environment, greatly enhancing the practical value and promotion potential of the technology.

[0019] Further, in step S3, the effective fracture surface area characterizing the fracture conductivity is obtained by inversion calculation using the fracture diagnostic model, wherein the fracture diagnostic model is constructed as a global transient analysis model. Step S3 includes: S301, setting a superimposed time variable and constructing a diagnostic chart with the superimposed time variable as the horizontal axis and the corresponding bottom hole flowing pressure or simulated bottom hole pressure as the vertical axis; S302, performing linear regression analysis on the data in the diagnostic chart to identify the linear flow characteristic segments and obtaining the absolute value of the slope of the regression line of the linear flow characteristic segments; S303, inverting and calculating the effective fracture surface area based on the absolute value of the slope; wherein the effective fracture surface area is constructed to be inversely proportional to the absolute value of the slope of the regression line.

[0020] By deeply integrating the Global Transient Analysis (UTA) diagnostic model with an automated data preprocessing pipeline, a leap has been made in crack status assessment, from "manual offline analysis months later" to "automatic online system evaluation at the hour / day level." This enables real-time closed-loop control, allowing for early detection of crack damage signs and timely intervention.

[0021] By introducing a "superimposed time variable," the complex variable production history is linearized. In the linear flow stage of fractured well production, the bottomhole flowing pressure (or pseudo-pressure) is linearly related to the square root of the superimposed time. The slope (m) of this linear relationship is inversely proportional to the effective fracture surface area of ​​the fracture system. Therefore, by automatically identifying the linear segment on the diagnostic chart and calculating its slope, the effective fracture surface area can be estimated in real time and quantitatively. This transforms the Precision Reservoir Engineering Theory (UTA), originally used for offline, post-hoc analysis, into a real-time diagnostic tool that can run online and automatically output key status parameters, achieving high-precision and quantitative perception of fracture health status.

[0022] Furthermore, when the target production well is an oil well, the vertical axis of the diagnostic map is the bottom hole flowing pressure, and the calculation of the effective fracture surface area is based on the following linear relationship: When the target production well is a gas well, the vertical axis of the diagnostic map is the simulated bottom-hole pressure based on actual gas, and the calculation of the effective fracture surface area is based on the following linear relationship: In the formula, The slope of the regression line is the absolute value. This is the crude oil volume coefficient. For fluid viscosity, For reservoir permeability, Porosity The overall compression coefficient is... The effective fracture surface area is T, and the formation temperature is T.

[0023] Further, step S4 includes: monitoring the changing trend of the effective crack surface area; when the monitored attenuation of the effective crack surface area exceeds a preset dynamic safety threshold, it is determined that the crack's flow conduction capacity is impaired, and an instruction to reduce the nozzle opening is generated; when the monitored effective crack surface area does not attenuate or the attenuation does not exceed the preset dynamic safety threshold, it is determined that the crack's flow conduction capacity is stable, and an instruction to maintain the current nozzle opening is generated; when the monitored effective crack surface area shows an upward trend, it is determined that the crack's flow conduction capacity is healthy, and an instruction to increase the current nozzle opening is generated. This quantifies the abstract crack "health status" as the changing trend of the effective crack surface area A relative to a certain benchmark value (safety threshold). By setting a reasonable attenuation tolerance limit, the system can automatically identify whether the crack is in a "damaged," "stable," or "healthy" state, and trigger corresponding protective (reducing nozzle opening), maintenance, or production-enhancing (increasing nozzle opening) control actions. This decision-making logic directly connects "state perception" and "control execution," realizing the automation and intelligence of production regulation. This enables the system to make optimal decisions autonomously based on the actual downhole conditions, just like "autopilot," thereby protecting the fracture system while dynamically exploring the production capacity limit of a single well and optimizing the production system in a personalized manner.

[0024] Furthermore, the dynamic safety threshold is set to the effective crack surface area obtained from the inversion before the last nozzle change. This uses a dynamically tracked benchmark, rather than a fixed value; that is, after each successful nozzle adjustment, the system records the effective crack surface area at the moment of adjustment as the new safety threshold. This allows the decision benchmark to be dynamically updated following the natural decay (or improvement) trend of the crack system, and makes the decision rules adaptive, avoiding potential system misjudgments that might occur due to using a fixed threshold (such as using the effective crack surface area at the initial high production level as the benchmark after the crack has already decayed significantly). This further improves the robustness and accuracy of the control system.

[0025] In the method of this application, the UTA model originally used for offline theoretical analysis is deeply integrated with adaptive data preprocessing (denoising, segmentation) and automated feature recognition (linear flow segment recognition, slope regression) algorithms, enabling it to process real-time data streams with noisy and variable working conditions on site, thereby outputting the effective crack surface area online and continuously.

[0026] Furthermore, this application proposes a two-stage denoising method that combines statistical thresholding (to eliminate gross errors) and adaptive window smoothing filtering (to filter out high-frequency noise), as well as an automatic segmentation algorithm based on time-varying derivatives to identify production change points. This combination is specifically designed to address the pain point of high noise in field data and frequent changes in operating conditions leading to the failure of traditional charts, ensuring the quality and applicability of subsequent UTA model input data.

[0027] The second aspect of this invention discloses a single-well production control system for fractured wells based on real-time diagnosis, used to implement the method disclosed in the first aspect of this application. The system includes: a data acquisition module configured to acquire single-well production data of a target production well in real time, the single-well production data including pressure data and production data; a data preprocessing module configured to preprocess the single-well production data; a diagnostic inversion module configured to obtain an effective fracture surface area characterizing fracture conductivity through inversion calculation based on the preprocessed single-well production data; and an intelligent decision-making module configured to generate and execute control commands for adjusting the nozzle opening of the target production well based on the real-time changing trend of the effective fracture surface area.

[0028] This solution is a system corresponding to the method. Its technical principle is to map the functional steps described in the method to specific hardware / software modules, constructing a physical entity system that achieves closed-loop control. Each module works collaboratively, realizing full-process automation from data acquisition to command execution.

[0029] More importantly, this application constructs an integrated system architecture of "virtual perception - real-time inversion - closed-loop control," which creatively integrates a virtual flow meter and pressure reconstruction module (addressing hardware deficiencies), a diagnostic inversion module integrating UTA (addressing state perception), and an intelligent decision-making module based on the effective crack surface area threshold rule (addressing automatic control). These three modules are interconnected, constructing a complete intelligent control system that does not rely on expensive hardware, can extract deep state information from conventional data, and can automatically act accordingly, thus realizing the implementation of the method.

[0030] Furthermore, the data acquisition module includes: a field data acquisition device for real-time acquisition of wellhead pressure and temperature data of the target production well; and a virtual flow meter module, which is configured to calculate and output the total equivalent production of oil, gas, and water phases based on the wellhead pressure and temperature data from the field data acquisition device, through a built-in wellbore multiphase flow physical model and data-driven model, and to calculate and acquire bottom hole flowing pressure data based on the wellhead pressure data and the total equivalent production.

[0031] Furthermore, the diagnostic inversion module has a built-in global transient analysis model for performing the following calculations: calculating the superimposed time variable based on production data, constructing a diagnostic chart with the superimposed time variable as the horizontal axis and the corresponding bottom hole flowing pressure or bottom hole pseudo-pressure as the vertical axis; performing linear regression analysis on the data in the diagnostic chart to identify the linear flow characteristic segment and obtaining the absolute value of the slope of the regression line of the linear flow characteristic segment; and inverting the calculation of the effective fracture surface area based on the absolute value of the slope.

[0032] Overall, this invention integrates virtual data perception, high-fidelity processing, real-time precision diagnosis, and intelligent decision execution, forming a complete "perception-analysis-decision-control" intelligent closed-loop system. It enables autonomous optimization management of the entire fracturing well production process, reduces reliance on professional human experience, and improves the level of intelligence in oilfield management.

[0033] Beneficial effects: This invention achieves continuous and reliable supply of key data at low cost through virtualized data acquisition and high-fidelity preprocessing; by using the effective fracture surface area as the core of real-time diagnosis, it achieves online and accurate perception of fracture health status; and finally forms a closed loop of "perception-diagnosis-decision-control", realizing adaptive production regulation of single wells, effectively delaying the decay of fracture conductivity and improving the final recovery rate.

[0034] The following describes in detail the method and system for controlling single-well production in fractured wells based on real-time diagnosis, with reference to the embodiments shown in the accompanying drawings and the reference numerals. Attached Figure Description

[0035] Figure 1 This is a flowchart of the steps in the single-well production control method for fracturing wells based on real-time diagnosis in this invention;

[0036] Figure 2 This is a schematic diagram comparing the effects of data preprocessing in this invention, where (A) is the original noisy data, (B) is the reconstructed signal after denoising, and (C) is the result of automatic flow segmentation based on the reconstructed signal.

[0037] Figure 3 This is a schematic diagram of the diagnostic principle based on the multiphase global transient analysis (UTA) model in this invention, showing the difference between the linear flow characteristic segment and the crack-damaged / nonlinear segment on the superimposed time-quasi-pressure diagnostic chart;

[0038] Figure 4 This is a schematic diagram illustrating the operating principle and data flow direction of the invention based on virtual flow meter and bottom well pressure reconstruction;

[0039] Figure 5 This is a schematic diagram of the architecture of the single-well production control system based on real-time diagnosis in this invention. Figure Labels

[0040] 10-Oil and gas well; 100-Field data acquisition device; 101-Pressure sensor; 102-Temperature sensor; 110-Virtual flow meter module; 200-Central processing system (edge ​​computing controller); 210-Data preprocessing module; 211-Pressure reconstruction unit; 212-Adaptive noise reduction unit; 213-Automatic segmentation unit; 220-Diagnostic inversion module; 221-Multiphase UTA model unit; 222-Fracture area calculation unit; 230-Intelligent decision-making module; 231-Health assessment unit; 232-Control command generation unit; 300-Field physical layer; 310-Intelligent nozzle actuator; 400-Remote monitoring terminal. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. 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. In addition, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0042] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0043] In this article, "single-well production data" refers to physical quantities that reflect the flow state of the wellbore and formation during the production process of oil and gas wells (especially unconventional oil and gas wells such as shale oil and shale gas). Specifically, this includes, but is not limited to: tubing / casing head pressure, bottom-hole pressure (BHP), fluid temperature, and instantaneous production of the three phases of oil, gas, and water. The bottom-hole pressure can be directly measured by a downhole pressure gauge or reconstructed from the wellhead pressure and a wellbore flow model.

[0044] The "real-time acquisition" or "real-time diagnosis" in this article does not refer solely to absolute synchronization at the millisecond or second level, but rather to the process of online data acquisition and calculation immediately after its generation. Its time resolution is adaptive: for drastically fluctuating wellhead pressure data, high-frequency acquisition (e.g., second-level) is preferred; while for gradually changing production data (e.g., tank level measurement), low-frequency acquisition (e.g., hourly-level) can be used. This invention, by introducing superimposed time variables, can handle such multi-frequency, non-equidistant time series data.

[0045] In this article, "virtual metering" refers to a software-based measurement technology. For operating conditions where expensive physical multiphase flow meters or separators are lacking, it is a method to indirectly calculate the flow rates of oil, gas, and water three-phase fluids by collecting easily accessible boundary conditions (such as wellhead pressure, temperature, and valve opening) and wellbore parameters, and using fluid dynamics mechanism models or data-driven models (such as neural networks).

[0046] The "fracture-matrix linear flow model" in this paper, also known as the fracture-formation linear flow model, refers to the flow state of fluid flowing vertically from a matrix with extremely low permeability to the fracture surface. Under this flow condition, the change in bottom hole flowing pressure exhibits a linear relationship with the square root of the stacking time, which is the physical basis for UTA diagnosis in this invention.

[0047] The "Multiphase Global Transient Analysis (UTA)" in this paper refers to an analytical method that integrates pressure transient analysis (PTA) and production transient analysis (RTA) into a single framework. This method introduces the "Convolution Time" variable, enabling it to handle various operating conditions such as continuous production, variable flow production, and shut-in pressure recovery within a unified coordinate system, thereby inverting formation parameters.

[0048] The term "effective crack surface area" in this article refers to... "Fracturing surface area" refers to the total area of ​​the fracture wall that actually participates in fluid flow and has effective conductivity within the fracture network formed by hydraulic fracturing. This parameter is a core indicator for evaluating fracturing effectiveness and monitoring fracture health (such as whether closure or embedding occurs).

[0049] This invention discloses a single-well production control method for fractured wells based on real-time diagnosis. This method aims to overcome the shortcomings of existing technologies, such as static control strategies, reliance on manual analysis leading to feedback lag, and poor quality of field multiphase flow data. Combined with... Figure 1As shown, the method includes the following steps: S1, acquiring single-well production data of the target production well in real time, the single-well production data including pressure data and production data; S2, preprocessing the single-well production data; S3, based on the preprocessed single-well production data, obtaining the effective fracture surface area characterizing fracture conductivity through inversion calculation; S4, generating and executing control commands for adjusting the nozzle opening of the target production well according to the real-time change trend of the effective fracture surface area.

[0050] Regarding step S1, the field data acquisition device 100 installed at the wellhead (see...) Figure 5 Completed. To capture transient pressure fluctuations that may occur during unconventional oil and gas well production (such as fluctuations caused by slug flow or fracture microvibrations), the data acquisition frequency is preferably set to high-frequency acquisition, such as 1Hz to 10Hz (i.e., 1 to 10 times per second), or hourly or minute-level acquisition. The acquired raw signals include wellhead pressure ( ) and fluid temperature ( ).

[0051] In a specific embodiment, step S1 includes: S101, acquiring the wellhead pressure data of the target production well in real time; S102, acquiring the total equivalent production as one of the production data, which can be obtained based on real-time measurement, or, based on the wellhead pressure data, calculating the continuous estimated production of the oil, gas, and water phases through a virtual flowmeter model, and then calculating the total equivalent production based on the continuous estimated production of the three phases; S103, acquiring the bottomhole flowing pressure data as one of the pressure data, which can be obtained based on real-time monitoring, or, based on the wellhead pressure data and the total equivalent production through back-calculation and reconstruction. In a more preferred embodiment, in step S101, the wellhead pressure data and wellhead temperature data of the target production well can be acquired in real time, and then in step S102, based on the wellhead pressure data and wellhead temperature data, the continuous estimated production of the oil, gas, and water phases can be calculated through a virtual flowmeter model, and then the total equivalent production can be calculated based on the continuous estimated production of the three phases, and then step S03 can be continued.

[0052] Furthermore, it should be specifically noted that in this invention, production data includes, but is not limited to, the instantaneous production of the three phases of oil / gas / water and the total equivalent production obtained by conversion; pressure data includes, but is not limited to, wellhead pressure and bottom hole flowing pressure (which can be measured or reconstructed by back calculation). S102 selects "total equivalent production" as "one of the production data" to accommodate two types of field paths: one is to obtain the total equivalent production by direct measurement, and the other is to estimate the three-phase production first through a virtual flow meter and then convert it to obtain the total equivalent production; other equivalent input forms such as three-phase production are not excluded.

[0053] Furthermore, considering the lack of physical metering equipment in actual operating conditions, step S102 includes a process of calculating multiphase output using a virtual flowmeter model. Simultaneously, to unify the contribution of different fluid phases to pressure, this step further utilizes the fluid PVT parameter (volume coefficient). , , The obtained three-phase production of oil, gas, and water is converted into the total equivalent production under formation conditions. For oil wells, the formula for calculating total equivalent production is: For gas wells, it is converted into total gas equivalent production.

[0054] It should be noted that the virtual flowmeter module described in this application belongs to the well-known software measurement technology in the art. Its function is to estimate the production of oil, gas, and water phases and the bottom hole flowing pressure based on easily obtainable wellhead monitoring data when physical multiphase flowmeters or downhole pressure gauges are lacking. This module itself does not constitute the inventive content of this invention, and the innovation of this invention does not lie in the construction or improvement of the virtual flowmeter model itself, but in the systematic integration of existing virtual metering methods with real-time diagnostic models and production control logic to achieve online perception and closed-loop control of fracture status in fracturing wells.

[0055] The specific implementation process of the virtual flow meter is described below:

[0056] See Figure 4 In view of the general lack of downhole pressure gauges and surface multiphase flow metering devices in shale oil and gas wells, this embodiment obtains key data through a virtual flow meter module 110 with integrated pressure reconstruction function.

[0057] Specifically, the virtual flow meter module receives real-time collected wellhead pressure ( ) and fluid temperature ( The module integrates a physical model and a neural network model for multiphase flow in the wellbore. The physical model uses the multiphase flow mechanism model in the wellbore to calculate the pressure gradient, liquid holdup, and friction loss during the gas-liquid two-phase or oil-gas-water three-phase flow process; the data-driven model is used to correct and compensate for the pressure drop characteristics under different flow conditions.

[0058] Based on the above fusion model, the system calculates the instantaneous production of the three phases of oil, gas, and water in real time. , , ), and output the back-calculated bottom hole flowing pressure ( The specific model form, parameter selection, and numerical solution method of the virtual flow meter module can all be implemented using multiphase flow virtual metering methods known in the art. This application does not limit this, and those skilled in the art can implement it according to the specific well type and working conditions, which will not be elaborated here.

[0059] Regarding step S2, the preprocessing steps for the single-well production data include a data cleaning step based on a two-stage adaptive denoising algorithm and a step of automatically identifying and segmenting the denoised data according to its flow regime. The data cleaning step includes: S201, setting a dynamic threshold based on the local variance and statistical distribution characteristics of the single-well production data to remove gross errors; S202, using an adaptive window smoothing filter combined with a median filter to perform secondary smoothing processing on the data after removing errors.

[0060] The steps for automatic flow pattern identification and segmentation include: S203, calculating the time-varying derivative and normalized rate of change of single-well production data; S204, when the time-varying derivative and normalized rate of change exceed the adaptive threshold, determining it as a change point in the production flow pattern; S205, automatically dividing the continuous time series into several independent flow analysis segments based on the change point, and removing invalid data segments with a time length less than a preset value.

[0061] See Figure 2 In one specific embodiment, in order to eliminate noise in the field data and extract the effective analysis segments, this application adopts the following processing flow:

[0062] 1) Steps of two-stage adaptive denoising:

[0063] Phase 1 (Gross Error Removal): For example... Figure 2 As shown in (A), the system calculates the local variance and statistical distribution characteristics of the data sequence. A statistical threshold is dynamically set (e.g., the 90th to 99th percentile of the data distribution). Data points that deviate from the local mean by more than this threshold (points marked with "×" in the figure) are identified as non-physical outliers and are either removed or replaced by interpolation.

[0064] Second stage (signal smoothing): such as Figure 2 As shown in (B), in order to preserve the true transient trend of pressure while filtering out high-frequency random noise (shown by the gray background in the figure), a second-level smoothing is performed using an adaptive window moving average filter (SMA) combined with a median filter. The so-called "adaptive window" means that the window length is not fixed, but automatically adjusted according to the fluctuation frequency of the data (e.g., dynamically changing between 0.1 days and 1.0 days), thereby obtaining a smooth feature data sequence (shown by the black solid line in the figure).

[0065] 2) Steps for automatic flow pattern identification and segmentation:

[0066] After the above two-stage denoising process, although the data curve becomes smoother (e.g. Figure 2(As shown in B), but the entire time series is still a continuous whole, including different operating conditions such as well opening and closing, and production adjustment. In order to meet the requirements of the UTA model for single linear flow analysis, this embodiment further segments the denoised data through an automatic segmentation unit 213. The specific steps are as follows:

[0067] Feature calculation: The system calculates the denoised stress data ( ) and total equivalent production data ( ) Calculate their first derivative with respect to time ( , (and normalized rate of change). This step is to capture the "rate" characteristic of data change.

[0068] Breakpoint identification: An adaptive threshold is set (e.g., based on the statistical mean or standard deviation of the absolute value sequence of derivatives). When the calculated absolute value of the derivative or the rate of change suddenly exceeds the threshold, the system determines that a "change in production flow pattern" has occurred at that moment (e.g., a sharp increase in the derivative usually corresponds to well opening / closing or a significant adjustment of the nozzle), and marks that moment as a breakpoint.

[0069] Sequence Segmentation and Cleaning: Based on the identified segmentation points, the continuous time series is segmented into several independent subsequences (i.e., flow analysis segments) on the time axis. To ensure the statistical significance of subsequent regression analysis, the system automatically checks the duration or number of data points of each subsequence. If the duration of a segment is less than a preset value (e.g., less than 12 hours or less than 50 data points), it is determined as an invalid data segment (usually a transitional and unstable noise segment) and is discarded. Only stable flow analysis segments are retained as input data for the subsequent construction of the fracture-matrix linear flow model.

[0070] Regarding step S3, the effective crack surface area characterizing the crack conductivity is obtained by inversion calculation through the crack diagnosis model, wherein the crack diagnosis model is constructed as a global transient analysis model.

[0071] In a specific embodiment, step S3 includes: S301, setting a superimposed time variable and constructing a diagnostic chart with the superimposed time variable as the horizontal axis and the corresponding bottom hole flowing pressure or simulated bottom hole pressure as the vertical axis; S302, performing linear regression analysis on the data in the diagnostic chart to identify linear flow characteristic segments and obtaining the absolute value of the slope of the regression line of the linear flow characteristic segments; S303, calculating the effective fracture surface area based on the absolute value of the slope; wherein the effective fracture surface area is constructed to be proportional to the absolute value of the slope of the regression line. Inversely proportional; furthermore, the global transient analysis model in engineering implementation consists of functional links such as superposition time calculation, diagnostic chart construction, linear flow feature identification and regression, and fracture surface area inversion calculation. It takes the pre-processed total equivalent production and bottom hole flowing pressure (or bottom hole pseudo pressure) as input, automatically outputs the effective fracture surface area, and corresponds to the multiphase UTA model unit 221 and fracture area calculation unit 222 in the system architecture, so that the effective fracture surface area can be automatically and continuously inverted and obtained under the condition of continuous production data in the field.

[0072] That is, the crack diagnosis model adopts the multiphase global transient analysis (UTA) model. This model linearizes the complex variable flow production history by introducing the "overlay time" variable. On the constructed diagnostic chart, linear flow characteristic segments are automatically identified, and the slope of the straight line is obtained through linear regression. According to UTA theory, this slope is related to the effective crack surface area ( The relationship is inversely proportional. Based on this relationship, the current value is calculated in real time. The physical mechanism is that, in the fracture-formation flow stage dominated by linear flow, the response of bottom hole flowing pressure (or bottom hole pseudo-pressure) to superposition time is determined by the size of the equivalent flow channel controlled by the fracture. Therefore, the regression slope of the linear segment on the diagnostic chart can characterize the change in the effective fracture surface area, thereby realizing the inversion calculation of the effective fracture surface area.

[0073] The multiphase global transient analysis (UTA) model used in this embodiment is based on the "fracture-formation linear flow model." This model describes the unique flow state of fluids in tight reservoirs, where fluid flows vertically from a matrix with extremely low permeability to the surface of artificial fractures. Under this specific flow condition, the change in bottomhole flowing pressure (or pseudo-pressure) exhibits a significant linear correlation with the square root of the convolution time. Based on this physical law, this embodiment introduces the concept of "convolution time." The variable ")" linearizes the complex variable flow production history, thereby achieving accurate inversion of the effective crack surface area, as detailed below:

[0074] First, the overlay time is calculated: the specific calculation formula is as follows. .in, It is the first Production at each time step ( ), For the first The time step This represents the current moment in the analysis. This variable transforms the complex history of variable flux into a single time variable through mathematical convolution.

[0075] Then perform diagnostic charting and inversion: See [link / reference] Figure 3 A diagnostic chart is constructed with the calculated "overlay time" as the horizontal axis and "bottom hole flowing pressure" (for oil wells) or "actual gas simulated pressure" (for gas wells) as the vertical axis.

[0076] Next, the linear flow feature segments are identified: such as Figure 3 As shown by the solid line, when the fluid flows linearly along the fractures towards the wellbore in the formation, the data points appear as a straight line on the chart. The system performs robust linear regression on this linear segment.

[0077] Then, slope extraction is performed: the absolute value of the slope of the regression line is obtained. According to UTA theory, this slope is related to the effective crack surface area. Inversely proportional.

[0078] Finally, parameter calculations are performed: the specific calculation formula for oil wells is as follows: The specific calculation formula for gas wells is as follows: Using the above formula, the system can calculate the effective crack surface area in real time and quantitatively. In the formula, The slope of the regression line is the absolute value. This is the crude oil volume coefficient. For fluid viscosity, For reservoir permeability, Porosity The overall compression coefficient is... The effective fracture surface area is T, and the formation temperature is T.

[0079] Regarding step S4, it includes: monitoring the changing trend of the effective crack surface area; when the monitored decrease in the effective crack surface area exceeds a preset dynamic safety threshold, it is determined that the crack flow conduction capacity is impaired, and an instruction to reduce the nozzle opening is generated; when the monitored effective crack surface area does not decrease or the decrease does not exceed the preset dynamic safety threshold, it is determined that the crack flow conduction capacity is stable, and an instruction to maintain the current nozzle opening is generated; when the monitored effective crack surface area shows an upward trend, it is determined that the crack flow conduction capacity is healthy, and an instruction to increase the current nozzle opening is generated.

[0080] That is, state identification is performed based on the parameters of the effective crack surface area calculated in step S3: such as Figure 3 As shown by the hollow point, when the data point deviates downward from the straight line (i.e., the slope) When the voltage is increased, it means that the voltage drop is greater in the same superposition time. This intuitively reflects... The reduction in conductivity indicates an impairment of the crack's conductivity. The system triggers the "protection mode" described in Example 1 by capturing this "deviation from linearity."

[0081] Specifically, the system compares the effective crack surface area obtained from the real-time inversion in step S3 with the historical baseline value. To accurately capture the dynamic changes in crack conductivity, this invention introduces a "safety threshold" as a judgment criterion. The safety threshold is selected as the effective crack surface area value before the last nozzle opening adjustment action. The system calculates in real-time the change in the current effective crack surface area relative to this safety threshold:

[0082] 1. If the current value remains stable or shows an upward trend compared to the safety threshold, the crack is determined to be in a "healthy state";

[0083] 2. If the current value decreases significantly compared to the safety threshold, and the decrease exceeds the preset tolerance limit (e.g., decrease exceeds 5%), the crack is determined to be in a state of "impaired flow conduction capacity".

[0084] In step S4, a control command is generated based on the evaluation results, and the production control device is driven to execute the control command to adjust the nozzle opening of the target production well. Based on the above health status determination results, the system generates the following specific control commands through an intelligent decision-making algorithm:

[0085] 1. For "impaired conductivity" (protection mode): To prevent stress-sensitive fracture closure, the system automatically generates a "reduce nozzle opening" command (e.g., reduce the opening by 1 / 64 inch). This operation aims to increase bottomhole flowing pressure by increasing wellhead back pressure, thereby reducing the effective stress acting on the fracture surface and restoring or maintaining the fracture's conductivity.

[0086] 2. For "Health Status" (Production Enhancement / Maintenance Mode): In order to maximize the production capacity of a single well while ensuring fracture health, the system generates instructions to "maintain the current opening" or "increase the nozzle opening" to explore the production capacity limit of the well under the current formation energy.

[0087] The present invention also discloses a single-well production control system for fractured wells based on real-time diagnosis, used to implement the method disclosed in the first aspect of this application, comprising: a data acquisition module configured to acquire single-well production data of a target production well in real time, the single-well production data including pressure data and production data; a data preprocessing module configured to preprocess the single-well production data; a diagnostic inversion module configured to obtain an effective fracture surface area characterizing fracture conductivity through inversion calculation based on the preprocessed single-well production data; and an intelligent decision-making module configured to generate and execute control commands for adjusting the nozzle opening of the target production well according to the real-time change trend of the effective fracture surface area.

[0088] This solution is a system corresponding to the method. Its technical principle is to map the functional steps described in the method to specific hardware / software modules, constructing a physical entity system that achieves closed-loop control. Each module works collaboratively, realizing full-process automation from data acquisition to command execution.

[0089] In a specific embodiment, the data acquisition module includes: a field data acquisition device for real-time acquisition of wellhead pressure and temperature data of the target production well; and a virtual flow meter module configured to calculate and output the total equivalent production of oil, gas, and water phases based on the wellhead pressure and temperature data from the field data acquisition device, using a built-in wellbore multiphase flow physical model and data-driven model, and to calculate and acquire bottom hole flowing pressure data based on the wellhead pressure data and the total equivalent production.

[0090] In a specific embodiment, the diagnostic inversion module has a built-in global transient analysis model for performing the following calculations: calculating the superimposed time variable based on production data, constructing a diagnostic chart with the superimposed time variable as the horizontal axis and the corresponding bottom hole flowing pressure or bottom hole pseudo-pressure as the vertical axis; performing linear regression analysis on the data in the diagnostic chart to identify the linear flow characteristic segment and obtaining the absolute value of the slope of the regression line of the linear flow characteristic segment; and calculating the effective fracture surface area based on the absolute value of the slope.

[0091] The specific implementation process of the system disclosed in this invention is as follows:

[0092] See Figure 5 The system mainly consists of three parts: the field physical layer 300, the central processing system 200, and the remote monitoring terminal 400.

[0093] Regarding the on-site physical layer 300, it directly interacts with the oil and gas well 10 and is responsible for the acquisition of physical signals and the execution of actions, as detailed below.

[0094] The field data acquisition device 100 includes a high-precision pressure sensor 101 (e.g., range 0-100 MPa, accuracy 0.1% FS) and a temperature sensor 102. These sensors are connected to the central processing system via explosion-proof cables and are responsible for acquiring raw physical signals (analog or digital).

[0095] The intelligent nozzle actuator 310 is installed downstream of the oil production tree valve. It comprises a specially designed throttle valve body and a servo motor drive unit. It can interpret digital instructions (such as Modbus RTU protocol instructions) from the central processing system, driving the valve core to make micron-level movements, thereby precisely controlling the flow area of ​​the fluid passage.

[0096] Regarding the Central Processing System 200 (edge ​​computing layer), it is deployed within an explosion-proof enclosure at the well site. The core hardware is preferably an edge computing controller (such as an industrial computer based on ARM architecture or a Raspberry Pi 4B or later). Its operating system can be Linux, and it contains multiple functional modules.

[0097] The virtual flow meter module 110 is configured to perform the flow and pressure calculation functions described in steps S102 and S103.

[0098] The data preprocessing module 210 includes a pressure reconstruction unit 211, an adaptive denoising unit 212, and an automatic segmentation unit 213. Its function is to transform the original high-frequency noise data into feature data suitable for model analysis.

[0099] The diagnostic inversion module 220 has a built-in crack-matrix linear flow model algorithm, including a multiphase UTA model unit 221 and a crack area calculation unit 222. This module acts as the "brain" of the system, responsible for executing the complex mathematical operations described in step S3.

[0100] The intelligent decision-making module 230 includes a health assessment unit 231 and a control instruction generation unit 232. It is responsible for executing the logical judgments described in Embodiment 1 and converting the decision results into hardware control signals.

[0101] Regarding the remote monitoring terminal 400 (cloud / user layer), located in the control center, it communicates bidirectionally with the central processing system 200 at the wellhead via a 4G / 5G / BeiDou satellite network. Users can view real-time UTA diagnostic diagrams (such as...) through this terminal. Figure 3 ),history The curve changes, and control can be remotely taken over in emergencies.

[0102] Overall, this invention integrates virtual data perception, high-fidelity processing, real-time precision diagnosis, and intelligent decision execution, forming a complete "perception-analysis-decision-control" intelligent closed-loop system. It enables autonomous optimization management of the entire fracturing well production process, reduces reliance on professional human experience, and improves the level of intelligence in oilfield management.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling the production of a single fracturing well based on real-time diagnosis, characterized in that, Includes the following steps: S1, real-time acquisition of single-well production data of the target production well, wherein the single-well production data includes pressure data and production data; S2, preprocess the single-well production data; S3, based on preprocessed single-well production data, the effective fracture surface area characterizing fracture conductivity is obtained through inversion calculation; S4. Based on the real-time change trend of the effective fracture surface area, generate and execute control commands for adjusting the nozzle opening of the target production well.

2. The method according to claim 1, characterized in that, Step S1 includes: S101, real-time acquisition of wellhead pressure data of the target production well; S102, obtain the total equivalent production as one of the production data, which can be obtained based on real-time measurement, or based on wellhead pressure data, calculate the continuous estimated production of oil, gas and water three phases through a virtual flow meter model, and then calculate the total equivalent production based on the continuous estimated production of the three phases. S103, Obtain bottom hole flowing pressure data as one of the pressure data, which can be obtained based on real-time monitoring, or obtained based on wellhead pressure data and total equivalent production through back calculation and reconstruction.

3. The method according to claim 1, characterized in that, The preprocessing steps for the single-well production data include a data cleaning step based on a two-stage adaptive denoising algorithm, which includes: S201 sets a dynamic threshold based on the local variance and statistical distribution characteristics of single-well production data to eliminate gross errors; S202 uses an adaptive window smoothing filter combined with a median filter to perform secondary smoothing on the data after removing errors.

4. The method according to claim 3, characterized in that, In the preprocessing steps of single-well production data, after the data cleaning step based on a two-stage adaptive denoising algorithm, a step of automatically identifying and segmenting the denoised data based on its flow regime is also included, which includes: S203, calculates the time-varying derivative and normalized rate of change of single-well production data; S204, when the time-varying derivative and normalized rate of change exceed the adaptive threshold, it is determined to be a change point in the production flow state; S205, based on the change point, the continuous time series is automatically divided into several independent flow analysis segments, and invalid data segments with a time length less than a preset value are removed.

5. The method according to claim 2, characterized in that, In step S3, the effective crack surface area characterizing the crack conductivity is obtained by inversion calculation through the crack diagnosis model, wherein the crack diagnosis model is constructed as a global transient analysis model. Step S3 includes: S301, Set the superimposed time variable to construct a diagnostic chart with the superimposed time variable as the horizontal axis and the corresponding bottom hole flowing pressure or bottom hole pseudo pressure as the vertical axis; S302, Perform linear regression analysis on the data in the diagnostic chart to identify the linear flow characteristic segments and obtain the absolute value of the slope of the regression line of the linear flow characteristic segments; S303, based on the absolute value of the slope, the effective crack surface area is calculated by inversion; wherein the effective crack surface area is constructed to be inversely proportional to the absolute value of the slope of the regression line.

6. The method according to claim 5, characterized in that, When the target production well is an oil well, the vertical axis of the diagnostic map is the bottom hole flowing pressure, and the calculation of the effective fracture surface area is based on the following linear relationship: ; When the target production well is a gas well, the vertical axis of the diagnostic map is the simulated bottom-hole pressure based on actual gas, and the calculation of the effective fracture surface area is based on the following linear relationship: ; In the formula, The slope of the regression line is the absolute value. This is the crude oil volume coefficient. For fluid viscosity, For reservoir permeability, Porosity The overall compression coefficient is... The effective fracture surface area is T, and the formation temperature is T.

7. The method according to claim 1, characterized in that, Step S4 includes: Monitor the changing trend of effective crack surface area; When the attenuation of the effective crack surface area exceeds the preset dynamic safety threshold, it is determined that the crack's flow conduction capacity is impaired, and an instruction to reduce the nozzle opening is generated. When the effective crack surface area is not detected to have decreased or the decrease rate does not exceed the preset dynamic safety threshold, it is determined that the crack flow conduction capacity is stable, and an instruction to maintain the current nozzle opening is generated. When the effective crack surface area is detected to be increasing, it is determined that the crack's flow conduction capacity is healthy, and an instruction to increase the current nozzle opening is generated.

8. The method according to claim 7, characterized in that, The dynamic safety threshold is set to the effective crack surface area obtained by inversion before the last nozzle change.

9. A single-well production control system for fractured wells based on real-time diagnostics, used to implement the method according to any one of claims 1-8, characterized in that, include: The data acquisition module is configured to acquire single-well production data of the target production well in real time, wherein the single-well production data includes pressure data and production data; The data preprocessing module is configured to preprocess the single-well production data; The diagnostic inversion module is constructed based on preprocessed single-well production data and obtains the effective fracture surface area characterizing fracture conductivity through inversion calculation. The intelligent decision-making module is configured to generate and execute control commands for adjusting the nozzle opening of the target production well based on the real-time changing trend of the effective fracture surface area.

10. The system according to claim 9, characterized in that, The data acquisition module includes: The on-site data acquisition device acquires real-time wellhead pressure and temperature data of the target production well; The virtual flow meter module is designed to calculate and output the total equivalent production of oil, gas, and water phases based on wellhead pressure and temperature data from a field data acquisition device, using a built-in wellbore multiphase flow physical model and data-driven model. It also calculates and obtains bottom hole flowing pressure data based on wellhead pressure data and total equivalent production.

11. The system according to claim 10, characterized in that, The diagnostic inversion module has a built-in global transient analysis model for performing the following calculations: Based on production data, superimposed time variables are calculated, and a diagnostic chart is constructed with the superimposed time variables as the horizontal axis and the corresponding bottom hole flowing pressure or bottom hole pseudo pressure as the vertical axis. Linear regression analysis is performed on the data in the diagnostic chart to identify linear flow characteristic segments and obtain the absolute value of the slope of the regression line of the linear flow characteristic segments. The effective crack surface area is calculated by inversion based on the absolute value of the slope.

Images