Underground oil-water separation intelligent adaptation system and method based on real-time data driving

Abstract

The invention discloses an underground oil-water separation intelligent adaptation system and method based on real-time data driving. The underground oil-water separation intelligent adaptation system comprises a three-way intelligent valve which is controlled by a central data processing system instruction and is connected with a lower main pipeline, a standby oil conveying pipe and a separation section main pipeline; five high-precision electric guide rails are arranged in a hollow hole structure in the middle of the wheel disc base, so that the modular separation pipeline is quickly switched along a preset direction; the modular separation pipeline comprises five crude oil pipelines, the lower end of any pipeline is connected with the separation section main pipeline, and the upper end of any pipeline is connected with the upper main pipeline; and the standby oil conveying pipe is connected between the intelligent valve and the upper main pipeline. A dynamic weight distribution mechanism is constructed through fusion modeling of multiple linear regression and a grey prediction algorithm, accurate matching of a separation module and an oil-water ratio is achieved, a separation pipeline and an electric wheel disc base are modularized, the maintenance cost is reduced by 50% or above, the separation efficiency is improved by 90% or above through an established full-process closed-loop optimization system, and the service life of equipment is prolonged to 8000 hours.

Description

Technical Field

[0001] The present invention belongs to the technical field of oilfield downhole intelligent equipment, and in particular relates to a downhole oil-water separation intelligent adaptation system and method driven by real-time data. Background Art

[0002] As oilfield development enters a high-water-cut phase (water cut >80%), pressures such as development costs, surface investment, water treatment investment, and environmental pollution are becoming increasingly prominent. To alleviate surface treatment pressures and reduce treatment equipment investment and operating costs, single-well production and injection technologies are being widely adopted. Downhole oil-water separation technologies, such as axial-flow hydrocyclones, have become a research hotspot due to their advantages, such as the lack of surface treatment and low energy consumption. Within the field of downhole oil-water separation technology, axial-flow liquid-liquid hydrocyclones have attracted attention due to their compact structure, small size, and light weight. Axial-inlet hydrocyclones rely on a guide vane structure at the inlet to achieve internal fluid diversion, thereby reducing turbulence intensity at the inlet section of the cyclone tube, enhancing flow field stability, and minimizing the probability of oil droplet breakup. This effectively minimizes pressure drop and energy consumption (Zhang Chunying. Structural Design and Performance Research of Axial-Flow Guide Vane Hydrocyclones [D]. China University of Petroleum (East China), 2021).

[0003] The separation effect of a cyclone separator is affected by many factors, including structural parameters, dispersed phase content, medium viscosity, density difference, inlet pressure, and flow rate (Ke Wenqi, Li Jianping, Wu Jiayi, Yang Lihong. Current status and prospects of research on same-well production and injection technology [J]. Mining Engineering, 2020, 8(3): 245-256.). Its technical bottlenecks mainly exist in the following three aspects. First, the cyclone separator has poor adaptability to working conditions. For traditional separators with fixed structural parameters, it can only reach a peak efficiency of about 85% at a specific oil-water ratio. When the oil-water ratio fluctuates to 20% or 70%, the efficiency drops sharply to below 60%. Second, the cyclone separator has insufficient tolerance to the environment. In high temperature, high pressure, or highly corrosive environments, the separator is prone to scaling and aging. However, due to the limitations of underground operating conditions, the replacement cycle is as long as 6-12 months. Furthermore, traditional cyclone separators generally have a technical bottleneck of lack of intelligence, that is, they lack a dynamic control mechanism driven by real-time data and rely solely on manual experience adjustment, resulting in a delayed response. Summary of the Invention

[0004] The purpose of the present invention is to provide a downhole oil-water separation intelligent adaptation system and method based on real-time data drive.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] An adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithms, comprising a lower main pipeline, a three-way intelligent valve, a spare oil pipeline, a wheel base, a modular separation pipeline, an upper main pipeline, and a separation section main pipeline;

[0007] The three-way intelligent valve is controlled by instructions from the central data processing system. The lower connecting port is connected to the lower main pipeline, the side connecting port is connected to the spare oil pipeline, and the upper connecting port is connected to the separation section main pipeline. The middle of the wheel base is a hollow hole structure, through which the crude oil pipeline passes. The hollow hole structure has five built-in high-precision electric guide rails, which support the rapid switching of modular separation pipelines along preset directions. The modular separation pipeline includes five crude oil pipelines of different specifications. The lower end of any one pipeline is connected to the separation section main pipeline, and the upper end is connected to the upper main pipeline. The spare oil pipeline is connected between the three-way intelligent valve and the upper main pipeline, and is used to allow the crude oil in the main pipeline to flow through the spare pipeline when the separator is replaced.

[0008] Furthermore, a lower end integrated probe and an upper end integrated probe are installed on the lower main pipeline and the upper main pipeline respectively; the integrated probes include a temperature sensor, a density meter, a viscometer, and an oil-water ratio detector.

[0009] Furthermore, the wheel base is a circular base, and the hollow structure in the middle is a mounting hole for an electric guide rail set at every 72° along the center point. A total of 5 mounting holes are connected in an "*" shape, and 5 high-precision electric guide rails are installed in the holes.

[0010] Furthermore, the five crude oil pipelines of the modular separation pipeline have differentiated designs in terms of inner diameter, blade inclination angle and overflow port diameter.

[0011] A method for an adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm, the specific steps are as follows:

[0012] Step 1: The integrated probe collects data every 0.1 seconds and preprocesses the data using the FTRL online learning algorithm;

[0013] Step 2: Upload all data to the central data processing system via optical fiber;

[0014] Step 3: Use multiple linear regression and grey prediction method to perform weighted averaging of data and calculate the comprehensive weight ω k ;

[0015] Step 4: Calculate the final adaptation ratio according to the adaptation formula to obtain the optimal oil-water ratio threshold boundary and real-time oil-water ratio of the separation section;

[0016] Step 5: The central system calculates the adaptability of the five crude oil pipelines of the modular separation pipeline, S1 to S5, and selects S minFor the corresponding pipeline, if the current module efficiency E is less than 85% or the oil-water ratio deviates from the adaptation range by more than 5%, the pipeline replacement instruction is triggered;

[0017] Step 6: Switch the three-way intelligent valve to the backup pipeline, and the wheel base drives the target crude oil pipeline into the center position to achieve pipeline switching;

[0018] Step 7: Data is transmitted back, model parameters are updated and stored in the database.

[0019] Furthermore, the step 1 adopts the FTRL online learning algorithm to filter the noise data and normalize it to the interval [0, 1].

[0020] Furthermore, the step 3 uses a multiple linear regression model to perform dynamic weight calculation:

[0021] Y=0.28X1+0.26X2+0.24X3+0.22X4

[0022] Among them, X1 is the oil-water ratio, X2 is the temperature, X3 is the viscosity, X4 is the density, and Y is the weight value;

[0023] Conduct grey relational analysis:

[0024]

[0025] Among them, Δ i is the sequence difference, Z is the weight value;

[0026] Calculate the comprehensive weight using the following formula:

[0027] ω k =0.6Y+0.4Z

[0028] Among them, ω k is the comprehensive weight value.

[0029] Furthermore, the calculation formula of the fitness formula in step 4 is:

[0030]

[0031] Among them, a k 、b k Both are the threshold boundaries of the optimal oil-water ratio in the separation section; R is the real-time oil-water ratio.

[0032] Furthermore, the separation efficiency feedback formula for updating the model parameters in step 7 is:

[0033]

[0034] Where E is the separation efficiency; C iis the volume fraction of the oil phase at the inlet; F is the split ratio, which refers to the ratio of the flow rate entering the overflow port after separation to the total flow rate at the inlet; E j To simplify the efficiency, it is obtained from the instrument calibration and reflects the theoretical efficiency upper limit of the separator under ideal conditions.

[0035] A computer-readable storage medium stores a computer program / instruction, which, when executed by a processor, implements the steps of an adaptive downhole oil-water separation method integrating multiple linear regression and grey prediction algorithms.

[0036] The beneficial effects of the present invention are:

[0037] (1) Significantly improved separation efficiency. The system can not only operate stably within a wide range of oil-water ratios from 20% to 95%, with an average separation efficiency of 92% (an 18% increase over traditional fixed separators), but also through real-time data acquisition and a dynamic weight model (multiple linear regression + gray prediction), the system updates the adaptation parameters every 0.1 seconds, ensuring that the separation efficiency fluctuation is less than ±2%.

[0038] (2) Maintenance costs are significantly reduced. Remote monitoring and fault diagnosis are achieved through the human-machine interface, which can reduce the frequency of on-site operations and save labor costs. The modular separation pipeline can be switched within 30 seconds through the electric wheel base, which improves maintenance efficiency compared to traditional manual replacement (taking 2-4 hours).

[0039] (3) Comprehensive improvement in safety and reliability. The system sets multiple safety thresholds (such as oil-water ratio deviation > 5%, temperature > 160°C, and pressure > 40 MPa) and initiates the emergency switching procedure within 0.5 seconds after the alarm is triggered, reducing the accident rate.

[0040] (4) Significant environmental benefits. The system's energy-saving design (reduced pressure drop and optimized energy consumption) can reduce annual carbon emissions per well, meeting green oilfield development standards. In addition, efficient separation reduces the water content of produced fluids to below 5%, reducing surface water treatment volume and annual wastewater discharge.

[0041] (5) Intelligent and adaptive advantages. By separating efficiency feedback and dynamic weight updates, the system can optimize model parameters every 24 hours and gradually reduce prediction errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a structural schematic diagram of the present invention.

[0043] Figure 2 This is a flow chart of the system of the present invention.

[0044] In the figure: 1. Lower end integrated probe, 2. Lower end main pipeline, 3. Three-way intelligent valve, 4. Spare oil pipeline, 5. Wheel base, 6. Modular separation pipeline, 7. Upper end main pipeline, 8. Upper end integrated probe, 9. Sliding track and storage hole, 10. Separation section main pipeline. DETAILED DESCRIPTION

[0045] The present invention will be further described below with reference to the accompanying drawings.

[0046] This invention describes a downhole intelligent oil-water separation adaptation system driven by real-time data. By integrating multiple linear regression with a gray prediction algorithm, a dynamic weight allocation mechanism can be constructed to achieve precise matching of the separation module and the oil-water ratio (with an error of less than 3%). The modular separation pipeline and electric wheel base design simultaneously support rapid switching of downhole separation sections (taking less than 30 seconds), reducing maintenance costs by more than 50%. On this basis, a full-process closed-loop optimization system is established, which improves separation efficiency to over 90% through real-time feedback (sampling frequency 1Hz) and extends equipment life to 8,000 hours.

[0047] The overall structure mainly consists of a lower integrated probe 1, a lower main pipeline 2, a three-way intelligent valve 3, a spare oil pipeline 4, a wheel base 5, a modular separation pipeline 6, an upper main pipeline 7, an upper integrated probe 8, a sliding track and storage hole 9, and a separation section main pipeline 10. It should be noted that the three-way intelligent valve 3 connects the lower main pipeline 2, the spare pipeline 4, and the separation section main pipeline 7, and is controlled by the central data processing system. When the separator is replaced, the valve leading to the separation section main pipeline 10 is closed, and the valve leading to the spare pipeline is opened, and the crude oil flows into the upper main pipeline through the spare pipeline; after the replacement is completed, the crude oil is introduced into the separation section main pipeline 10 in the same way. In addition, there is a sliding track on the wheel base 5, and the modular separation pipeline 6 is fixed on the sliding track. The separation pipeline at the edge of the wheel base 5 is moved into or out of the main pipeline system via the sliding track to realize the replacement of the separation pipeline.

[0048] This system consists of three core modules: data acquisition module, execution module and control and optimization module.

[0049] For the data acquisition module, its high-precision integrated probe can be integrated with a variety of high-precision integrated probes with different functions, including: (1) temperature sensor: range -20℃~200℃, accuracy ±0.3℃ (using Pt100 platinum resistance); (2) density meter: based on vibration principle, resolution 0.05g / cm 3(3) Viscometer: Rotary measurement, range 1-500mPa·s, repeatability error <1%; (4) Oil-water ratio detector: Microwave phase method, real-time detection error <1.5%. Through the coordinated measurement of multiple probes, it is possible to detect crude oil temperature, flow rate, pressure, oil-water ratio, and kinematic viscosity data. At the same time, the data acquisition module uses an anti-interference transmission system (using armored optical fiber (temperature resistant to 200℃) and RS485 dual redundant communication, transmission delay <20ms, bit error rate <10-6) to perform terminal processing on the data.

[0050] For the execution module, the modular separation pipe material can be selected from Hastelloy C276, which has the characteristics of high temperature resistance and corrosion resistance. The pipe can be selected in five specifications, corresponding to the oil-water ratio range: 20%-35% (No. 1), 35%-50% (No. 2), 50%-65% (No. 3), 65%-80% (No. 4), and 80%-95% (No. 5). The pipe structure parameters adopt a differentiated design and can include the following designs: inner diameter 30mm / 40mm / 50mm, blade inclination angle 15° / 30° / 45°, overflow port diameter 10mm / 15mm / 20mm. The wheel base adopts a ring structure.

[0051] The 1.5m outer diameter houses five high-precision electric guide rails (positioning accuracy ±0.1mm), supporting switching of the separation pipe along 0°, 72°, 144°, 216°, and 288° positions. The drive motor is an explosion-proof servo motor (IP68 protection rating), with a designed torque of 50 N·m and a switching speed of 0.5 m / s. Furthermore, the three-way intelligent valve is designed to withstand a pressure rating of 60 MPa and a leakage rate of less than 0.01%. It is electromagnetically actuated and supports remote control from a central system.

[0052] For the control and optimization module, the FTRL (Follow-the-Regularized-Leader) online learning algorithm is first used in the data preprocessing stage to filter the noise data (such as temperature mutation >10℃ / s) and normalize it to the [0,1] range.

[0053] Then the dynamic weight calculation is performed, and the formula of the multiple linear regression model is as follows:

[0054] Y=0.28X1+0.26X2+0.24X3+0.22X4

[0055] Among them, X1 is the oil-water ratio, X2 is the temperature, X3 is the viscosity, X4 is the density, and Y is the weight value.

[0056] Carry out grey relational analysis, the specific formula is as follows:

[0057]

[0058] Among them, Δ i is the sequence difference, and Z is the weight value.

[0059] Calculate the comprehensive weight using the following formula:

[0060] ω k =0.6Y+0.4Z

[0061] Among them, ω k is the comprehensive weight value.

[0062] Adaptation decision, select the optimal module (the one with the smallest S value) based on the fitness formula, the expression is as follows:

[0063]

[0064] Among them, a k 、b k Both are the threshold boundaries of the optimal oil-water ratio in the separation section; R is the real-time oil-water ratio.

[0065] For example, assuming that the optimal oil-water ratio of separation section 1 is between 60% and 75%, and the viscosity adaptability is S 1，1 , the comprehensive weight ratio of viscosity is 0.3. At this time, the real-time oil-water ratio R=68%. According to the formula:

[0066] S 1,1 =0.3×(|60%-68%|+|75%-68%|)=4.5.

[0067] Calculate the temperature adaptability S by analogy 1，2 , density adaptation S 1，3 , oil-water ratio adaptation S 1，4 .

[0068] S1=S 1,1 +S 1,2 +S 1,3 +S 1,4

[0069] S1 is the total fitness of separation section 1. Similarly, after calculating the total fitness of the remaining separation sections, the one with the smallest S value is selected, which is the optimal separation section under the oil-water ratio.

[0070] The self-learning mechanism dynamically adjusts model parameters based on separation efficiency feedback and updates weight coefficients daily. The separation efficiency feedback formula is as follows:

[0071]

[0072] Where E is the separation efficiency; C i is the volume fraction of the oil phase at the inlet; F is the split ratio, which refers to the ratio of the flow rate entering the overflow port after separation to the total flow rate at the inlet; Ej To simplify the efficiency, it is obtained from the instrument calibration and reflects the theoretical efficiency upper limit of the separator under ideal conditions.

[0073] For the human-computer interaction interface, the separation efficiency curve, pipeline life progress bar, alarm information (oil-water ratio exceeds the limit, temperature is too high, etc.) can be displayed in real time, and manual intervention mode is supported, which can force the switching of separation modules or adjust the weight ratio.

[0074] The workflow of this system is as follows: first, data collection and transmission are carried out. The integrated probe collects data every 0.1 seconds and uploads the data to the central data processing system through optical fiber. Then, the weighted average of various data is performed by multivariate linear regression and gray prediction method, and the final adaptation ratio is calculated according to the adaptation formula. Finally, dynamic adaptation decision is made. The central system calculates the adaptation degree S1 to S5 of each module and selects S min If the current module efficiency (E) of the corresponding pipeline is less than 85% or the oil-water ratio deviates from the adaptation range by more than 5%, a replacement command is triggered. The three-way valve then switches to the backup pipeline, and the wheel base drives the target module into the main pipeline, completing the module switch. After the switch is complete, crude oil flows through the new module, and the old module is moved to the backup position for maintenance. Finally, closed-loop optimization is performed, which means that data is transmitted back after separation, and model parameters are updated and stored in a database (capacity 10TB, storage period ≥ 5 years).

[0075] See also Figure 2 The specific implementation plan for the central data processing system is as follows: Taking an offshore oilfield in the Bohai Sea (well depth 3200m, pressure 38MPa, oil-water ratio fluctuation 25%-75%) as an example, five separation modules were first installed, with specific parameters shown in the table below. After the probes were calibrated, the system was started. Real-time monitoring and decision-making were then carried out. At a specific moment, the parameters monitored yielded an oil-water ratio of 68%, a temperature of 135°C, a viscosity of 75mPa·s, and a density of 0.89g / cm 3 The central system calculated S3 = 0.32 and S4 = 0.18 (minimum), so the system switched to module 4. The module switch and verification process then proceeded, with the wheel base driving module 4 into the main pipeline. After the switch, separation efficiency increased from 78% to 93%. Finally, the system implemented long-term optimization, automatically updating weight coefficients daily. After 30 days, the model prediction error dropped from an initial 5% to 2.3%.

[0076]

[0077] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. An adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm, characterized by: It comprises a lower main pipeline (2), a three-way intelligent valve (3), a spare oil pipeline (4), a wheel base (5), a modular separation pipeline (6), an upper main pipeline (7), and a separation section main pipeline (10); The three-way intelligent valve (3) is controlled by instructions from a central data processing system, with a lower connection port connected to a lower main pipeline (2), a side connection port connected to a spare oil pipeline (4), and an upper connection port connected to a separation section main pipeline (10); the middle of the wheel base (5) is a hollow hole structure, through which the crude oil pipeline passes, and the hollow hole structure is built with five high-precision electric guide rails (9) to support the modular separation pipeline (6) to quickly switch along a preset direction; the modular separation pipeline (6) includes five crude oil pipelines of different specifications, the lower end of any one pipeline being connected to the separation section main pipeline (10), and the upper end being connected to the upper main pipeline (7); the spare oil pipeline (4) is connected between the three-way intelligent valve (3) and the upper main pipeline (7), and is used for allowing the crude oil in the main pipeline to flow through the spare pipeline when the separator is replaced.

2. The adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm according to claim 1, characterized in that: A lower end integrated probe (1) and an upper end integrated probe (8) are installed on the lower main pipeline (2) and the upper main pipeline (7), respectively; the integrated probes include a temperature sensor, a density meter, a viscometer, and an oil-water ratio detector.

3. The adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm according to claim 1 is characterized in that: The wheel base (5) is a circular base, and the hollow structure in the middle is a mounting hole for an electric guide rail (9) arranged at intervals of 72 degrees along the center point. A total of five mounting holes are connected to form an "*" shape, and five high-precision electric guide rails (9) are installed in the holes.

4. The adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm according to claim 1, characterized in that: The five crude oil pipelines of the modular separation pipeline (6) are designed with different inner diameters, blade inclination angles and overflow port diameters.

5. The method for an adaptive downhole oil-water separation system integrating multiple linear regression and grey prediction algorithm according to any one of claims 1 to 4, characterized in that: The specific steps are as follows: Step 1: The integrated probe collects data every 0.1 seconds and preprocesses the data using the FTRL online learning algorithm; Step 2: Upload all data to the central data processing system via optical fiber; Step 3: Use multiple linear regression and grey prediction method to perform weighted averaging of data and calculate the comprehensive weight ω k ; Step 4: Calculate the final adaptation ratio according to the adaptation formula to obtain the optimal oil-water ratio threshold boundary and real-time oil-water ratio of the separation section; Step 5: The central system calculates the adaptability S1 to S5 of the five crude oil pipelines of the modular separation pipeline (6), and selects S min For the corresponding pipeline, if the current module efficiency E is less than 85% or the oil-water ratio deviates from the adaptation range by more than 5%, the pipeline replacement instruction is triggered; Step 6: The three-way intelligent valve (3) is switched to the backup pipeline (4), and the wheel base (5) drives the target crude oil pipeline into the center position to realize pipeline switching; Step 7: Data is transmitted back, model parameters are updated and stored in the database.

6. The adaptive downhole oil-water separation method integrating multiple linear regression and grey prediction algorithm according to claim 5 is characterized in that: The step 1 uses the FTRL online learning algorithm to filter the noise data and normalize it to the [0, 1] interval.

7. The adaptive downhole oil-water separation method integrating multiple linear regression and grey prediction algorithm according to claim 5 is characterized in that: Step 3 uses a multiple linear regression model to perform dynamic weight calculation: Y=0.28X1+0.26X2+0.24X3+0.22X4 Among them, X1 is the oil-water ratio, X2 is the temperature, X3 is the viscosity, X4 is the density, and Y is the weight value; Conduct grey relational analysis: Among them, Δ i is the sequence difference, Z is the weight value; Calculate the comprehensive weight using the following formula: oh k =0.6Y+0.4Z Among them, ω k is the comprehensive weight value.

8. The adaptive downhole oil-water separation method integrating multiple linear regression and grey prediction algorithm according to claim 5 is characterized in that: The calculation formula of the fitness degree in step 4 is: Among them, a k 、b k Both are the threshold boundaries of the optimal oil-water ratio in the separation section; R is the real-time oil-water ratio.

9. The adaptive downhole oil-water separation method integrating multiple linear regression and grey prediction algorithm according to claim 5 is characterized in that: The separation efficiency feedback formula for updating the model parameters in step 7 is: Where E is the separation efficiency; C i is the volume fraction of the oil phase at the inlet; F is the split ratio, which refers to the ratio of the flow rate entering the overflow port after separation to the total flow rate at the inlet; E j To simplify the efficiency, it is obtained from the instrument calibration and reflects the theoretical efficiency upper limit of the separator under ideal conditions.

10. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 5 to 9 are implemented.

Images