An edge-computing-based downhole ultrasonic atomization control system and method

Abstract

The application discloses a kind of based on edge computing's downhole ultrasonic atomization control system and method, it is related to coalbed gas drainage and downhole intelligent control technical field.The method includes: obtaining multi-source state data and constructing the composite jam state vector consisting of coal dust jam component, water lock component and coupling component;According to jam evolution stage, determine target atomization state reversely and map into ultrasonic and liquid supply control parameters;Control downhole ultrasonic atomization main body to generate medicament droplet and adjust gas lift parameter synchronously;Collect downhole near field and wellhead far field double feedback data to determine jam removal effective degree and control parameter switching or re-adjust;Update control parameter mapping relationship through online learning and carry out predictive anti-jam intervention according to jam trend.The application realizes accurate identification of coal dust and water lock composite damage, adaptive control of atomization state, collaborative control of ultrasonic and gas lift, double feedback evaluation and predictive anti-jam, effectively improves coalbed gas well jam removal efficiency and long-term drainage stability.

Description

Technical Field

[0001] This invention relates to the fields of coalbed methane drainage, reservoir unblocking, and downhole intelligent control technology, specifically to a downhole ultrasonic atomization control system and method based on edge computing. Background Technology

[0002] Coalbed methane (CBM), as an important component of unconventional natural gas, relies primarily on long-term drainage and depressurization to gradually desorb adsorbed methane within the coal seam and achieve stable gas production. However, during long-term drainage and production in CBM wells, factors such as coal body structure fragmentation, coal dust migration, hydraulic disturbance, and repeated changes in wellbore fluids can easily lead to coal dust deposition at perforation orifices, fracture inlets, and throats. Coal dust deposition reduces the effective flow area of ​​the orifices and throats, increases fluid seepage resistance, and further induces fluid retention at the bottom of the well.

[0003] On the other hand, coalbed methane reservoirs typically have strong hydrophilicity. When liquid remains trapped inside pores and fractures for extended periods, a liquid film adheres to the pore wall surface, forming a waterlock phenomenon that further reduces methane permeability. Coal dust blockage and waterlock damage are not independent but rather exhibit a clear coupling effect. Coal dust deposition narrows the pore throat, making liquid drainage difficult and thus exacerbating waterlock formation. Conversely, waterlock, by causing liquid retention, reduces the fluid's ability to carry coal dust, leading to continued coal dust deposition and ultimately resulting in a combined reservoir damage caused by the combined effects of coal dust blockage and waterlock.

[0004] Existing coalbed methane well unblocking technologies mainly include mechanical flushing, chemical unblocking, high-pressure injection, gas lift and return flow, and ultrasonic-assisted unblocking.

[0005] The existing technology has the following main shortcomings: First, it cannot distinguish between coal dust blockage and water lock-dominated damage states, making it difficult to accurately locate the main controlling factors of blockage; Second, it cannot dynamically adjust droplet size and ultrasonic parameters according to the degree of blockage, resulting in low efficiency of reagent action; Third, ultrasonic unblocking and gas lift backflow are usually independent of each other, lacking a coordinated control mechanism; Fourth, it only monitors the operating status of the equipment and cannot evaluate whether the reservoir has truly restored its conductivity; Fifth, it lacks a learning mechanism for the historical response of a single well, making it difficult to achieve long-term anti-blockage.

[0006] Therefore, it is necessary to propose a downhole ultrasonic atomization control system and method based on edge computing. Summary of the Invention

[0007] The purpose of this invention is to provide a downhole ultrasonic atomization control system and method based on edge computing, so as to solve the technical problems existing in the coalbed methane well unblocking technology, such as inaccurate identification of blockage status, inability to adaptively adjust ultrasonic parameters, lack of coordination between ultrasonic and gas lift, lack of reservoir response evaluation and online learning mechanism.

[0008] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0009] A closed-loop control method for downhole ultrasonic atomization based on edge computing includes the following steps:

[0010] Acquire multi-source status data during the coalbed methane well drainage process. The multi-source status data includes at least downhole pressure, temperature, atomized liquid supply status, ultrasonic transducer electrical response, ultrasonic action response, gas production flow rate, return liquid volume, water content, and return coal powder status.

[0011] Based on the multi-source state data, a composite blockage state vector of the target reservoir segment is constructed. The composite blockage state vector consists of a first state component representing the degree of coal powder deposition, a second state component representing the degree of water lock, and a third state component representing the degree of coupling promotion between coal powder and water lock. The current blockage evolution stage is determined based on the position of the composite blockage state vector in the state space.

[0012] The target atomization state is determined in reverse based on the blockage evolution stage, and the target atomization state is mapped to the ultrasonic control parameters and fluid supply control parameters of the downhole ultrasonic atomization body.

[0013] The downhole ultrasonic atomizing body is controlled to generate drug droplets according to the ultrasonic control parameters and liquid supply control parameters, and the gas lift control parameters are adjusted synchronously during the droplet transport process so that the drug droplets generate ultrasonic cavitation effect in the target reservoir section that matches the current complex blockage state;

[0014] During the unblocking process, downhole near-field feedback data and wellhead far-field feedback data are collected. The effectiveness of the ultrasonic excitation and atomization process is determined based on the downhole near-field feedback data, and the effectiveness of the unblocking product return is determined based on the wellhead far-field feedback data.

[0015] The current unblocking effectiveness is determined based on the corresponding change relationship between the downhole near-field feedback data and the wellhead far-field feedback data, and the control parameters are switched or readjusted based on the unblocking effectiveness.

[0016] The system associates and stores the current complex blockage state, current control parameters, and current deblocking effectiveness. It updates the control parameter mapping relationship through online learning and performs early anti-blocking intervention based on the changing trend of the complex blockage state, forming a closed-loop control system of deblocking, backflow, recovery, and predictive anti-blocking.

[0017] A closed-loop control system for downhole ultrasonic atomization based on edge computing, comprising:

[0018] The ground edge control unit is used to receive multi-source state data, construct a composite blockage state vector, generate ultrasonic control parameters, liquid supply control parameters and gas lift control parameters, and perform online learning and updates.

[0019] The liquid preparation and supply module is connected to the surface edge control unit and is used to provide unblocking liquid with preset concentration and supply status to the well according to the liquid supply control parameters.

[0020] The downhole ultrasonic atomizing unit is inserted into the target reservoir section through a tubing string and is used to convert the unblocking agent into agent droplets in the target atomized state according to the ultrasonic control parameters.

[0021] The downhole near-field monitoring module is connected to the downhole ultrasonic atomizing body and is used to collect the downhole near-field feedback data;

[0022] The wellhead far-field monitoring module is used to collect the wellhead far-field feedback data;

[0023] The gas lift return module is used to provide flow-carrying effect to the downhole according to the gas lift control parameters in order to regulate droplet transport and carry unblocking products back to the well.

[0024] The surface edge control unit calculates the unblocking effectiveness based on the downhole near-field feedback data and the wellhead far-field feedback data, and then jointly corrects the ultrasonic control parameters, fluid supply control parameters, and gas lift control parameters based on the unblocking effectiveness, and then updates the unblocking result to the online learning.

[0025] This invention offers the following advantages: It enables accurate identification of combined damage caused by pulverized coal and water lock. Unlike traditional methods that rely solely on pressure or flow rate, this invention constructs a composite blockage state vector consisting of a pulverized coal blockage component, a water lock component, and a coupling component. By determining the position and directional derivatives in the three-dimensional state space, it identifies the blockage evolution stage, significantly improving the accuracy and timeliness of blockage state identification.

[0026] This invention achieves adaptive reverse control of the atomization state. Based on different clogging states, the target droplet state is determined through optimized inversion, enabling coordinated changes in ultrasonic frequency, power, duty cycle, and liquid supply parameters. This transforms the output from fixed parameters to a customized atomization strategy based on the clogging state, improving the efficiency of drug entry into the pore throat and its action.

[0027] This invention achieves coordinated control of ultrasound and air lift. Air lift simultaneously performs the functions of droplet transport and pulverized coal return, enabling the unblocking process to be completed continuously, avoiding secondary deposition of unblocking products, and improving drainage efficiency.

[0028] This invention establishes a dual-feedback evaluation mechanism. The effectiveness of unblocking is jointly evaluated through cross-correlation analysis of the downhole equipment response (near field) and the wellhead reservoir response (far field), avoiding erroneous judgments based solely on equipment operating status and improving the reliability of control decisions.

[0029] This invention continuously accumulates the correspondence between different blockage states and control parameters through a recursive estimation method, enabling subsequent unblocking to gradually form the optimal control strategy for a single well and achieve the self-evolution of control parameters. Through real-time monitoring of the blockage evolution trend, low-intensity intermittent anti-blocking is initiated in advance when the compound blockage has not yet reached the unblocking threshold but its growth trend reaches the preset anti-blocking condition, thereby reducing the risk of long-term production capacity decline in coalbed methane wells. Attached Figure Description

[0030] Figure 1 This is a flowchart of the method steps of the present invention;

[0031] Figure 2 This is a schematic diagram of the overall structure of the downhole ultrasonic atomization closed-loop control system of the present invention;

[0032] Figure 3 This is a schematic diagram of the three-dimensional state space of the composite blockage state vector of the present invention;

[0033] Figure 4 This is a structural diagram of the main structure of the downhole ultrasonic atomization system of the present invention;

[0034] Figure 5 This is a timing diagram illustrating the online learning and predictive anti-blocking control of the present invention. Detailed Implementation

[0035] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0036] like Figures 1-5 As shown, the present invention provides a closed-loop control system for downhole ultrasonic atomization based on edge computing, including a surface edge control unit, a liquid preparation and supply module, a downhole ultrasonic atomization body, a downhole near-field monitoring module, a wellhead far-field monitoring module, and a gas lift return module.

[0037] The surface edge control unit is located on the wellhead surface and communicates with each module downhole via cable or fiber optic cable. The surface edge control unit has a built-in edge computing processor used to receive multi-source state data, construct a composite blockage state vector, generate ultrasonic control parameters, fluid supply control parameters, and gas lift control parameters, and perform online learning and updates.

[0038] The liquid preparation and supply module is connected to the surface edge control unit and is installed at the wellhead. It is used to provide the unblocking solution with a preset concentration and supply status to the downhole ultrasonic atomizing body according to the liquid supply control parameters. The liquid preparation and supply module includes a storage tank, a liquid preparation pump, a flow regulating valve, and a liquid level sensor.

[0039] The downhole ultrasonic atomizer is inserted into the target reservoir section (i.e., the perforation section) through the tubing string. It is used to convert the unblocking agent into agent droplets in the target atomized state according to the ultrasonic control parameters.

[0040] The downhole near-field monitoring module is connected to the downhole ultrasonic atomizing body and is located inside or near the downhole ultrasonic atomizing body. It is used to collect near-field feedback data that reflects ultrasonic excitation, atomization status and downhole action response.

[0041] The wellhead far-field monitoring module is installed on the wellhead surface to collect far-field feedback data reflecting the reservoir conductivity recovery and unblocking product backflow status of the target reservoir section.

[0042] The gas lift return module is installed on the wellhead surface and connected to the tubing string. It is used to provide flow carrying effect to the downhole according to the gas lift control parameters output by the surface edge control unit, so as to regulate the transport of chemical droplets and carry back coal powder, water and residual chemicals generated during unblocking.

[0043] Construction and identification of composite congestion state vectors: Construction of composite congestion state vectors.

[0044] Define a composite congestion state vector ,in:

[0045] First state component The degree of coal powder deposition is characterized by the cumulative effect of changes in reservoir pressure differential and the concentration of backflowed coal powder.

[0046]

[0047] in, The reservoir pressure differential change (unit: MPa) is acquired in real time by downhole pressure sensors. The concentration of pulverized coal returned to the wellhead (unit: mg / L) is obtained through sampling and analysis of the wellhead return fluid or online turbidity monitoring. and These are non-negative weighting coefficients, calibrated through indoor experiments based on the geological characteristics of the target reservoir. This is the start time of the sampling process.

[0048] Second state component The degree of water lock is characterized by the degree of gas production flow rate reduction and the degree of water saturation anomaly.

[0049]

[0050] in, The current gas production flow rate (unit: m³ / d) is collected in real time by a wellhead flow meter. The initial gas production flow rate (unit: m³ / d) is the average value taken during the stable gas production stage in the initial stage of drainage and extraction. The water saturation anomaly index (dimensionless, value range 0~1) is obtained through downhole resistivity logging or capacitive water cut monitoring. and These are non-negative weighting coefficients.

[0051] Third state component The degree of coupling promotion between pulverized coal blockage and water lock is characterized and determined based on the synergistic enhancement effect between the first and second state components:

[0052]

[0053] Wherein, λ is the non-negative coupling coefficient, which is calibrated through historical data analysis or indoor simulation experiments; This indicates taking the smaller of the two values, reflecting the "weakest link" constraint characteristic of the coupling effect; This indicates that the larger of the two rates of change is taken, reflecting the dynamic acceleration characteristics of the coupling effect.

[0054] Based on the norm of the composite congestion state vector in the three-dimensional state space The directional derivative determines the current stage of congestion evolution:

[0055] when and When the preset pulverized coal dominance threshold is exceeded, it is determined to be a pulverized coal dominance blockage state;

[0056] when and When the preset water lock dominance threshold is exceeded, it is determined to be a water lock dominance blockage state;

[0057] when and Synchronous growth and Exceeding the preset coupling threshold When this occurs, it is determined to be a complex blockage state;

[0058] when When the value continues to decrease and all components are below the recovery threshold, it is determined to be in a state of unblocking and recovery.

[0059] In practical applications, the calibration methods for weighting coefficients and thresholds are as follows: Select 3-5 completed production wells in the target block, collect their historical drainage data (including gas production, water production, pulverized coal concentration, downhole pressure, etc.), and combine this with permeability damage data obtained from core experiments to determine the weighting coefficients and thresholds through least squares fitting. The optimal value for the coupling threshold ξ is determined by statistically analyzing the values ​​of each production well when a significant production capacity decline occurs. The average value is taken as the initial threshold, and then adjusted online in subsequent applications based on the actual unblocking effect.

[0060] Inverse calculation of the target atomization state: The target atomization state is determined by inverse calculation based on the composite clogging state vector. The target atomization state includes at least the median particle size of the target droplets. Droplet size distribution dispersion Liquid supply per unit time and the effective transport rate of target droplets .

[0061] Inverse calculus is achieved by solving an optimization problem that minimizes the mapping deviation between the atomized state and the clogging state.

[0062]

[0063] Where S is the atomization state vector. Let M be the target atomization state vector, M be the positive mapping operator between the atomization state and the blockage state (obtained by fitting indoor atomization-unblocking experimental data), and Ω be the state space that the downhole atomization body can realize (limited by the transducer frequency response range, maximum driving power, and liquid supply system capacity).

[0064] Ultrasonic control parameters include at least the ultrasonic drive frequency. Drive power and incentive duty cycle The liquid supply control parameters should include at least the liquid supply flow rate. and liquid supply pressure .

[0065] By coordinating changes in at least two of the ultrasonic drive frequency, drive power, excitation duty cycle, and liquid supply flow rate and pressure, the actual atomization state can be made to approximate the target atomization state. For example, when coal powder is determined to be the main cause of blockage, larger droplet sizes are preferred. (15-25μm) to enhance impact stripping effect; when water-locked blockage is determined to be the main cause, smaller droplet size (15-25μm) is preferred. (5-10 μm) to improve specific surface area and chemical reaction efficiency; when determined to be complex blockage, medium particle size ( (with a thickness of 10-15μm) and combined with high driving power to achieve a combined effect of liquid film disruption and coal powder desorption.

[0066] Optimization of ultrasonic control parameters: Ultrasonic control parameters are determined using a combination of trial excitation and response evaluation. Multiple different ultrasonic excitation combinations are applied within a preset parameter space, and the response data corresponding to each ultrasonic excitation combination is obtained.

[0067] Specifically, in the preset parameter space Apply N different combinations of ultrasonic excitations internally (i=1,2,…,N), each parameter combination includes different driving frequencies, driving power, and excitation duty cycles. Obtain the corresponding response data for each group; the response data should at least include the transducer impedance change. RMS value of drive current Amplitude of pressure fluctuation in atomizing chamber and incremental response of pulverized coal discharge .

[0068] Constructing the cost function :

[0069]

[0070] in, Here is the transducer reference impedance (unit: Ω, obtained under clean water conditions), and ϵ is a small quantity to prevent the denominator from being zero (taken as...). ), , , , Normalized weight coefficients and satisfying The principle for determining the weighting coefficients is: assign higher weights to response indicators that have a greater impact on congestion relief. Generally, it is recommended to... =0.2, =0.15, =0.35, =0.3, and the specific value can be fine-tuned according to the characteristics of the target reservoir.

[0071] The ultrasonic parameter combination that can simultaneously satisfy the conditions for atomization stability, cavitation intensity, and reservoir response improvement is selected. Furthermore, the parameter combination in which each individual indicator is not lower than the preset minimum threshold will be selected as the working parameter for the current unblocking cycle.

[0072] In practical applications, the value of N is determined based on the dimension and accuracy requirements of the parameter space, and N=5-9 is generally recommended. The time interval between trial excitations should not be too long. Response data should be collected after each set of parameters has been running for 3-5 minutes. The entire optimization process should be completed within 30-60 minutes to avoid unnecessary disturbance to the reservoir caused by prolonged trials.

[0073] Dynamic coordinated control of gas lift parameters: Gas lift control parameters are dynamically determined based on droplet transport status and unblocking product return status. Gas lift control parameters include at least the gas lift injection flow rate. Heqi Lift Injection Pressure .

[0074] During the ultrasonic atomization stage, based on the downhole pressure The droplet retention and settling trend in the wellbore is determined by the fluid supply status and current gas lift flow rate. A droplet retention and settling trend coefficient is defined. :

[0075] )

[0076] in The droplet settling velocity (unit: m / s) is calculated from the droplet size and fluid properties according to Stokes' law. This is the gas lift flow rate (unit: m / s), calculated from the gas lift injection flow rate and wellbore cross-sectional area. When When the velocity is ≥1, it indicates that the droplet settling velocity has reached or exceeded the airlift flow rate, and the droplets cannot be effectively transported to the target perforation section. At this point, the airlift injection flow rate should be increased until... <1.

[0077] During the unblocking and backflow stage, the concentration of backflowed pulverized coal is used as a guide. , backflow volume and gas production flow rate Determine the extent of unblocking product formation :

[0078]

[0079] in , , Non-negative weights and , and These are the preset maximum allowable values ​​for the concentration of pulverized coal and the volume of return fluid, respectively (determined based on the wellbore's fluid carrying capacity and surface treatment capacity). Adjust the size of the air lift injection flow rate: A larger value indicates more unblocking products, requiring a larger airlift flow rate to ensure timely backflow.

[0080] Therefore, the airlift process simultaneously performs the dual functions of droplet transport and unblocking product return.

[0081] Acquisition and correlation analysis of dual feedback data: Downhole near-field feedback data should include at least the driving voltage of the ultrasonic transducer. Drive current Operating impedance Atomizing chamber pressure and downhole pressure fluctuations At least two of the following. These data are collected in real time by a downhole near-field monitoring module, with a sampling frequency of no less than 100Hz.

[0082] Wellhead far-field feedback data should include at least changes in gas production flow rate. , Casing pressure change , backflow volume Moisture content index and the concentration of returned pulverized coal At least two of the following. These data are collected in real time by the wellhead far-field monitoring module, with a sampling frequency of no less than 1Hz.

[0083] Correlation analysis was performed between downhole near-field feedback data and wellhead far-field feedback data according to their time correspondence, and the cross-correlation function between the two was calculated:

[0084]

[0085] in, This is the downhole near-field feedback vector. For the wellhead far-field feedback vector, It is a cross-correlation function. and These are the standard deviations of near-field feedback and far-field feedback, respectively.

[0086] When the cross-correlation function The first threshold was not exceeded within the preset time delay interval. If the ultrasonic effect is not effectively transmitted to the reservoir (i.e., the equipment is idling or energy is dissipated within the wellbore), the current ultrasonic control parameters should be reduced and the target atomization state redefined. Exceeding the second threshold When the ultrasonic action is determined to be effectively transmitted to the reservoir and produce a corresponding reservoir response, the current ultrasonic atomization parameters are maintained and the unblocking effect evaluation stage is entered.

[0087] In practical applications, time delay interval The range of values ​​is determined based on the well depth and fluid propagation velocity. Generally, it is 0-30 minutes for shallow wells (<1000m), 0-60 minutes for medium-deep wells (1000-2000m), and 0-120 minutes for deep wells (>2000m). Generally, a value of 0.3-0.4 is used. The value is generally set to 0.6-0.7, but the specific value should be adjusted according to the noise level and signal quality at the site.

[0088] Comprehensive evaluation of unblocking effectiveness: The unblocking effectiveness Φ(t)∈[0,1] is determined jointly based on the changes in reservoir response before and after ultrasonic treatment.

[0089]

[0090] in, To calculate the incremental recovery of gas production flow (unit: m³ / d), the difference between the stable gas production flow after ultrasonic treatment and the gas production flow before ultrasonic treatment is taken. The reservoir pressure recovery increment (unit: MPa) is obtained through downhole pressure gauge monitoring. The change in pulverized coal concentration during return flow (unit: mg / L) is calculated as the difference between the peak concentration and the background concentration. The change in water content is a dimensionless quantity, representing the difference in water saturation index before and after the action. This represents the change in transducer operating impedance (unit: Ω), calculated as the difference in operating impedance before and after operation. The superscript "max" indicates the preset maximum allowable change in each physical quantity (preset based on reservoir properties and equipment capabilities). to Normalized weights and satisfying .

[0091] The effectiveness level is divided according to the value of the decongestion relief effectiveness:

[0092] when When this happens, it is determined that the blockage has not been effectively resolved;

[0093] when When this occurs, it is determined that the surface blockage has been cleared;

[0094] when At that time, it was determined that the main blockage had been cleared;

[0095] when At that time, it was determined that the reservoir diversion was fully restored.

[0096] When it is determined that the surface blockage has been cleared but the reservoir flow recovery has not reached the target state, maintain gas lift backflow and restart ultrasonic atomization; when it is determined that the main blockage has been cleared or the reservoir flow has recovered, stop continuous ultrasonic action and switch to intermittent anti-blockage control.

[0097] In practical applications, Generally, a value of 0.2-0.3 is used. Generally, a value of 0.5-0.6 is used. The weight is typically set between 0.8 and 0.9. to The values ​​reflect the relative importance of different indicators. Generally, restoring gas production flow is the ultimate goal, and it is recommended... The maximum value is set (0.30-0.35), and the weights of the remaining indicators are allocated according to the actual situation.

[0098] Online learning is achieved by constructing a set of triples representing state, parameters, and responses. At the end of each deblocking cycle, the current composite congestion state vector is... Current control parameter combination and the corresponding effectiveness of unblocking Associative storage as triples .

[0099] A recursive estimation method is used to update the control parameter mapping. The state-parameter-response triplet set is defined. The least squares recursive estimation is used to update the mapping parameter matrix W:

[0100]

[0101]

[0102]

[0103] in, Let be the mapping weight matrix for the k-th unblocking cycle. For Kalman gain, Here is the covariance matrix, and μ is the forgetting factor (ranging from 0.9 to 0.99; the smaller the value, the faster the historical data is forgotten and the stronger the ability to adapt to new working conditions). The matrix is ​​the identity matrix. Through the above recursive estimation, the control parameters for subsequent unblocking cycles are adjusted based on the optimal effectiveness corresponding to the same type of complex congestion state in historical cycles.

[0104] The blockage trend is determined based on the rate of change of the combined blockage state over multiple consecutive drainage cycles. (Rate of change) Define the total rate of change:

[0105]

[0106] When the norm of the composite congestion state vector Below the decongestion threshold (That is, it has not yet reached the point where it needs to be unblocked) and the total rate of change Achieve preset anti-blocking conditions If the congestion trend is determined to develop rapidly, a low-intensity intermittent anti-congestion state is initiated in advance. The initiation time of low-intensity intermittent anti-congestion... and duration The system is dynamically determined based on historical congestion response data—specifically, by querying historical databases and comparing them with current data. and The three most similar historical records are used as a reference value for the current blocking strategy by taking the weighted average of their blocking start time and duration, so that the blocking control trigger time precedes the capacity decline time.

[0107] In practical applications, The calibration method is as follows: statistically analyze the historical production wells in the target block from their normal state to the point where significant production capacity decline (gas production decreases by more than 20%). The changing pattern is determined by taking the average of the 7 days prior to the occurrence of capacity decline. As The initial value is set and then adjusted based on the actual anti-blocking effect during application.

[0108] The downhole ultrasonic atomizer body includes a cylindrical hollow shell, an atomization chamber disposed inside the cylindrical hollow shell, a liquid input channel communicating with the atomization chamber, a piezoelectric ceramic ultrasonic transducer assembly disposed around the atomization chamber, and a mist exhaust channel communicating with the atomization chamber.

[0109] The cylindrical, perforated shell is made of corrosion-resistant stainless steel, with perforated structures on the side walls to allow the atomized reagent droplets to exit smoothly and enter the annulus. A centralizing and positioning structure, employing either an elastic centralizer or a rigid centralizer block, is installed on the outer side of the cylindrical, perforated shell to position the downhole ultrasonic atomizer at the perforation location in the target reservoir section, ensuring that the droplets accurately act on the target perforation section.

[0110] The piezoelectric ceramic ultrasonic transducer operates based on the frequency conversion and power control signals output by the ground edge control unit, changing the ultrasonic drive frequency. Drive power and incentive duty cycle Adjusting the state of the reagent droplets. The piezoelectric ceramic ultrasonic transducer uses a multi-piece piezoelectric ceramic stacked structure to obtain sufficient power output within the limited downhole space.

[0111] The liquid input channel is connected to the liquid supply pipeline inside the tubing, which in turn is connected to the liquid preparation and supply module on the ground. The mist exhaust channel is connected to the mist exhaust pipeline inside the tubing or directly to the annulus, ensuring that the atomized liquid droplets can be smoothly delivered to the target perforation section.

[0112] The ground edge control unit is also equipped with water shortage protection logic: when the liquid level of the liquid preparation and supply module is detected... Below the minimum working level Liquid supply flow rate Below minimum working flow Or the liquid supply pressure of the atomizing chamber Below minimum working pressure When any of the conditions is met, reduce or stop ultrasonic excitation to prevent the transducer from burning out and being damaged; after the liquid supply is restored, adjust the current composite blockage state vector accordingly. Re-determine the ultrasonic control parameters and fluid supply control parameters to ensure the unblocking effect;

[0113] The working principle is as follows:

[0114] Sensing Layer: Construction and Identification of Composite Blockage State Vector. The system acquires real-time data on pressure, temperature, gas production flow rate, flowback coal powder concentration, water cut status, and ultrasonic transducer electrical response from downhole and wellhead multi-source sensors. This transforms the previously unobservable reservoir blockage state into a quantifiable three-dimensional state vector. This vector consists of a coal powder deposition component, a water-locking component, and a coupling-promoting component between the two. By analyzing the position and directional derivatives in the three-dimensional state space, the dominant factors and evolution stage of the current blockage are accurately determined, providing a basis for subsequent control strategy formulation.

[0115] Decision-making level: Inverse calculation and parameter mapping of the target atomization state. Based on the identified blockage evolution stage, the system determines the target atomization state matching the current blockage state by solving an optimization inverse problem. For coal powder-dominated blockage, larger droplets are prioritized to enhance the impact stripping effect; for water-lock-dominated blockage, smaller droplets are prioritized to improve specific surface area and chemical reaction efficiency; for combined blockage, an intermediate state that balances both is adopted. The target atomization state is then mapped to specific control parameters such as ultrasonic drive frequency, drive power, excitation duty cycle, liquid supply flow rate, and liquid supply pressure. In addition, the system uses an optimization method combining trial excitation and response evaluation to screen the optimal operating parameters within a preset parameter space, ensuring that each unblocking operation starts in the optimal state.

[0116] Execution Layer: Coordinated Control of Ultrasonic Atomization and Gas Lift Backflow. The downhole ultrasonic atomization unit transforms the unblocking agent into target-state droplets according to control parameters issued by the decision-making layer. Simultaneously, the gas lift backflow module synchronously adjusts the gas lift injection flow rate and pressure. During the ultrasonic atomization stage, it carries the droplets towards the target perforation section; after the unblocking products are generated, it handles the backflow of coal dust, liquid, and residual agents. The combination of ultrasonic cavitation and chemical agent action synergistically disrupts the coal dust deposition structure and the water-locked liquid film, while the air lift's flow-carrying effect ensures the timely discharge of the disrupted products, preventing secondary deposition.

[0117] Feedback Layer: Dual-feedback closed-loop evaluation and adaptive parameter correction. The system simultaneously collects downhole near-field feedback data (reflecting the working status of the ultrasonic transducer, atomization chamber pressure, and downhole pressure fluctuations) and wellhead far-field feedback data (reflecting gas production flow recovery, changes in return coal powder concentration, and changes in water cut). By calculating the cross-correlation function between the two types of feedback data, it determines whether the ultrasonic energy is effectively transmitted to the reservoir and produces the expected reservoir flow recovery effect. Based on the comprehensive evaluation results of the unblocking effectiveness, the system automatically decides whether to maintain the current parameters to continue unblocking, switch to return mode, or readjust the control parameters, forming a dual-closed-loop control system with an inner loop (stable equipment status) and an outer loop (reservoir response convergence).

[0118] Learning Layer: Online Learning and Predictive Anti-blocking. After each unblocking cycle, the system stores the current complex blockage state, control parameter combination, and corresponding unblocking effectiveness as a triple. It continuously updates the control parameter mapping relationship through a recursive estimation method, allowing subsequent unblocking control parameter selection to be corrected based on the historical best response of a single well, thus achieving self-evolution of the control strategy. Simultaneously, by monitoring the rate of change of the complex blockage state, the system proactively initiates a low-intensity intermittent anti-blocking state when the blockage degree has not yet reached the unblocking threshold but its growth trend meets the preset anti-blocking conditions. This upgrades the traditional passive unblocking mode to a predictive active anti-blocking mode.

[0119] Overall, this invention, through the aforementioned five-order closed-loop control architecture, achieves a complete logical closed loop from reservoir state perception to control strategy generation, from ultrasonic and gas lift coordinated execution to dual feedback effect evaluation, and from historical experience learning to future trend prediction. This upgrades downhole ultrasonic atomization unblocking from an open-loop experience-driven mode to a closed-loop data-driven mode, effectively improving the accuracy, effectiveness, and long-term stability of unblocking complex blockages in coalbed methane wells.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed-loop control method for downhole ultrasonic atomization based on edge computing, characterized in that, Includes the following steps: S1. Obtain multi-source status data during the coalbed methane well drainage process. The multi-source status data includes at least downhole pressure, temperature, atomized liquid supply status, ultrasonic transducer electrical response, ultrasonic action response, gas production flow rate, return liquid volume, water content, and return coal powder status. S2. Construct a composite blockage state vector for the target reservoir segment based on the multi-source state data. The composite blockage state vector consists of a first state component representing the degree of coal powder deposition, a second state component representing the degree of water lock, and a third state component representing the degree of coupling promotion between coal powder and water lock. Determine the current blockage evolution stage based on the position of the composite blockage state vector in the state space. S3. Determine the target atomization state in reverse according to the blockage evolution stage, and map the target atomization state to the ultrasonic control parameters and fluid supply control parameters of the downhole ultrasonic atomization body; S4. Control the downhole ultrasonic atomizing body to generate drug droplets according to the ultrasonic control parameters and liquid supply control parameters, and simultaneously adjust the gas lift control parameters during the droplet transport process so that the drug droplets generate ultrasonic cavitation effect in the target reservoir section that matches the current complex blockage state. S5. During the unblocking process, collect downhole near-field feedback data and wellhead far-field feedback data. Determine the effectiveness of the ultrasonic excitation and atomization process based on the downhole near-field feedback data, and determine the effectiveness of the unblocking product return based on the wellhead far-field feedback data. S6. Determine the current unblocking effectiveness based on the corresponding change relationship between the downhole near-field feedback data and the wellhead far-field feedback data, and switch or readjust the control parameters based on the unblocking effectiveness. S7. The current complex blockage state, current control parameters, and current unblocking effectiveness are associated and stored. The control parameter mapping relationship is updated through online learning, and anti-blocking intervention is carried out in advance according to the changing trend of the complex blockage state, forming a closed-loop control of unblocking, backflow, recovery, and predictive anti-blocking.

2. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The method for constructing the composite blockage state vector includes: The composite blockage state vector is denoted as... ; Wherein, the first state component Determined based on the cumulative effect of reservoir pressure differential change and return coal powder concentration: Second state component Determined based on the degree of gas production flow rate decline and the degree of water saturation anomaly: The third state component Determined based on the synergistic enhancement effect between the first state component and the second state component: in, This represents the change in reservoir pressure differential. To determine the concentration of pulverized coal returned to the system, This represents the current gas production flow rate. This is the initial gas production flow rate. It is the water saturation anomaly index. , , , , All are non-negative weighting coefficients; Based on the norm of the composite congestion state vector in the three-dimensional state space The current blockage evolution stage is determined by its directional derivative: when the first state component is dominant, it is determined to be a pulverized coal-dominated state; when the second state component is dominant, it is determined to be a water-lock-dominated state; when the first state component and the second state component grow synchronously and the third state component exceeds a preset coupling threshold, it is determined to be a composite blockage state.

3. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The target atomization state is determined by inverse calculation based on the composite blockage state vector: The target atomization state includes at least the median droplet size, droplet size distribution dispersion, liquid supply per unit time, and effective transport rate of the target droplets. The reverse calculation is achieved by solving an optimization problem that minimizes the mapping deviation between the atomized state and the clogging state: Where S is the atomization state vector. Let be the target atomization state vector. This is a positive mapping operator between the atomized state and the blocked state. This represents the state space that the downhole atomization system can achieve. The ultrasonic control parameters include at least ultrasonic driving frequency, driving power and excitation duty cycle, and the liquid supply control parameters include at least liquid supply flow rate and liquid supply pressure. By coordinating changes in at least two of the ultrasonic drive frequency, drive power, excitation duty cycle, and liquid supply flow rate and liquid supply pressure, the actual atomization state can be made to approximate the target atomization state.

4. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The ultrasonic control parameters are determined by a combination of trial-and-error excitation and response evaluation. Multiple different ultrasonic excitation combinations are applied within a preset parameter space, and response data corresponding to each ultrasonic excitation combination are obtained. The response data includes at least the transducer impedance change, the effective value of the driving current, the pressure fluctuation amplitude of the atomization chamber, and the response increment of the returned coal powder. Establish a correspondence between ultrasonic excitation parameters and downhole action response based on the response data; The ultrasonic parameter combination that can simultaneously meet the conditions of atomization stability, cavitation intensity, and reservoir response improvement is selected, and the selected parameter combination is determined as the working parameter for the current unblocking cycle.

5. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The airlift control parameters are dynamically determined based on the droplet transport state and the return state of unblocking products: The air lift control parameters include at least the air lift injection flow rate and the air lift injection pressure; During the ultrasonic atomization stage, the retention and settling trend of droplets in the wellbore is judged based on the downhole pressure, fluid supply status and current gas lift flow rate. When the droplet settling velocity reaches or exceeds the gas lift flow rate, the gas lift injection flow rate is increased until the droplet settling velocity is lower than the gas lift flow rate. During the unblocking and return stage, the degree of unblocking product formation is determined based on the return coal powder concentration, return liquid volume, and gas production flow rate. The gas lift injection flow rate is then adjusted according to the degree of unblocking product formation, so that the gas lift process can simultaneously perform the dual functions of droplet transport and unblocking product return.

6. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The downhole near-field feedback data includes at least two of the following: the driving voltage, driving current, operating impedance of the ultrasonic transducer, atomization chamber pressure, and downhole pressure fluctuation. The wellhead far-field feedback data shall include at least two of the following: gas production flow rate change, casing pressure change, flowback fluid volume, water cut index, and flowback pulverized coal concentration. Correlation analysis was performed between downhole near-field feedback data and wellhead far-field feedback data according to their time correspondence, and the cross-correlation function between the two was calculated: in, This is the downhole near-field feedback vector. For the wellhead far-field feedback vector, It is a cross-correlation function. and These are the standard deviations of near-field feedback and far-field feedback, respectively. When the cross-correlation function does not exceed the first threshold within the preset time delay interval, it is determined that the ultrasonic action has not been effectively transmitted to the reservoir. At this time, the current ultrasonic control parameters are reduced and the target atomization state is redefined. When the cross-correlation function exceeds the second threshold, the current ultrasonic atomization parameters are maintained and the unblocking effect evaluation stage is entered.

7. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The effectiveness of unblocking is determined jointly based on the changes in reservoir response before and after ultrasonic treatment: The unblocking effectiveness Φ(t)∈[0,1] is determined by weighting the gas production flow recovery increment, reservoir pressure recovery increment, change in backflow coal powder concentration, change in water cut state, and change in transducer working impedance: in, To restore the incremental gas production flow, To increase the reservoir pressure recovery, This represents the change in the concentration of pulverized coal discharged. This represents the change in water content. This represents the change in the transducer's operating impedance; the superscript "max" indicates the preset maximum allowable change for each physical quantity. to Normalized weights and satisfying ； The effectiveness level is divided according to the value of the unblocking effectiveness: if it is lower than the first preset threshold, it is determined that the unblocking is not effective; if it reaches the first preset threshold and is lower than the second preset threshold, it is determined that the surface blockage is removed; if it reaches the second preset threshold and is lower than the third preset threshold, it is determined that the main blockage is removed; and if it reaches the third preset threshold, it is determined that the reservoir diversion is fully restored.

8. The closed-loop control method for downhole ultrasonic atomization based on edge computing according to claim 1, characterized in that, The online learning is achieved by constructing a set of state-parameter-response triples: After each unblocking cycle, the current composite blockage state vector, the current combination of control parameters, and the corresponding unblocking effectiveness are associated and stored as a triple; A recursive estimation method is used to update the control parameter mapping relationship, so that the control parameters for subsequent unblocking cycles are corrected based on the optimal effectiveness corresponding to the same type of complex blockage state in historical cycles. The blockage trend is determined based on the rate of change of the composite blockage state in multiple consecutive drainage cycles, wherein the rate of change is the sum of the rates of change of the first state component, the second state component, and the third state component. When the norm of the composite congestion state vector is lower than the unblocking threshold and the rate of change reaches the preset anti-congestion condition, the low-intensity intermittent anti-congestion state is activated in advance. The activation time and duration of the low-intensity intermittent anti-congestion state are dynamically determined based on the historical unblocking response.

9. A closed-loop control system for downhole ultrasonic atomization based on edge computing, used to execute the method as described in any one of claims 1 to 8, characterized in that, include: The ground edge control unit is used to receive multi-source state data, construct a composite blockage state vector, generate ultrasonic control parameters, liquid supply control parameters and gas lift control parameters, and perform online learning and updates. The liquid preparation and supply module is connected to the surface edge control unit and is used to provide unblocking liquid with preset concentration and supply status to the well according to the liquid supply control parameters. The downhole ultrasonic atomizing unit is inserted into the target reservoir section through a tubing string and is used to convert the unblocking agent into agent droplets in the target atomized state according to the ultrasonic control parameters. The downhole near-field monitoring module is connected to the downhole ultrasonic atomizing body and is used to collect the downhole near-field feedback data; The wellhead far-field monitoring module is used to collect the wellhead far-field feedback data; The gas lift return module is used to provide flow-carrying effect to the downhole according to the gas lift control parameters in order to regulate droplet transport and carry unblocking products back to the well. The surface edge control unit calculates the unblocking effectiveness based on the downhole near-field feedback data and the wellhead far-field feedback data, and then jointly corrects the ultrasonic control parameters, fluid supply control parameters, and gas lift control parameters based on the unblocking effectiveness, and then updates the unblocking result to the online learning.

10. A closed-loop control system for downhole ultrasonic atomization based on edge computing according to claim 9, characterized in that, The downhole ultrasonic atomizing body includes a cylindrical hollow shell, an atomizing cavity disposed inside the cylindrical hollow shell, a liquid input channel communicating with the atomizing cavity, a piezoelectric ceramic ultrasonic transducer disposed on the periphery of the atomizing cavity, and a mist exhaust channel communicating with the atomizing cavity. A centering and positioning structure is provided on the outside of the cylindrical hollow shell to position the downhole ultrasonic atomizing body at the perforation position of the target reservoir section; The piezoelectric ceramic ultrasonic transducer works according to the frequency conversion signal, power control signal and duty cycle control signal included in the ultrasonic control parameters, and adjusts the state of the drug droplets by changing the ultrasonic driving frequency, driving power and excitation duty cycle. The ground edge control unit is also equipped with a water shortage protection logic: when any of the following conditions are met, the liquid level of the liquid preparation and supply module is lower than the minimum working liquid level, the liquid supply flow rate is lower than the minimum working flow rate, or the liquid supply pressure of the atomizing chamber is lower than the minimum working pressure, the ultrasonic excitation is reduced or stopped; after the liquid supply state is restored, the ultrasonic control parameters and liquid supply control parameters are re-determined according to the current composite blockage state vector.

Images