Automatic gas lift control method for shale gas well liquid loading

Abstract

This invention provides an automatic gas lift control method for shale gas wells with accumulated liquid, relating to the field of shale gas extraction technology. The method involves real-time acquisition of production parameter data from the shale gas well using wellhead sensors; based on the production parameter data, liquid accumulation identification logic is used to determine the real-time liquid accumulation status of the gas well; when the real-time liquid accumulation status is determined to be either a liquid accumulation risk state or an already liquidated state, effective production data under the liquid-free state is retrieved from a liquid-free data pool, and a multi-model fusion algorithm is used to calculate the comprehensive critical liquid-carrying flow rate; based on the comprehensive critical liquid-carrying flow rate, a multi-objective optimization algorithm aiming to maximize gas injection efficiency is used to calculate the optimal gas injection rate; based on the optimal gas injection rate, a valve opening control command is generated, and the control command drives the gas injection valve to actuate, obtaining the gas lift control result. This invention solves the problems of reliance on manual labor, low control accuracy, and slow response in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of shale gas extraction technology, and in particular to an automatic gas lift control method for liquid accumulation in shale gas wells. Background Technology

[0002] As shale gas wells enter the later stages of production, the formation energy decreases, gas production capacity weakens, and the wellhead's fluid-carrying capacity becomes insufficient, making it prone to fluid accumulation at the bottom of the well. Fluid accumulation increases bottom-hole back pressure, reduces gas well productivity, and in severe cases, can even lead to well shutdown. Therefore, it is necessary to inject high-pressure gas into the well using a gas lift process to enhance fluid-carrying capacity and maintain stable production.

[0003] In the existing technology, the judgment of the timing of gas lift and the determination of the gas lift volume rely entirely on the manual experience of the on-site operators: the operators need to regularly inspect the wellhead parameters, judge whether there is a risk of liquid accumulation based on experience, and then adjust the opening of the gas injection valve to control the gas injection volume based on experience. This method has the following significant defects: (1) Judgment lag: manual inspection cannot monitor the liquid accumulation status in real time, and often gas lift is only started when the liquid accumulation has affected production or there is a risk of production stoppage, resulting in loss of gas well production capacity; (2) Low gas injection accuracy: lacking scientific model support, excessive gas injection volume will easily cause gas waste, while insufficient gas injection volume will not be able to effectively discharge the liquid accumulation, making it difficult to achieve the optimal gas lift effect; (3) Reliance on manual experience: the standardization of operation is greatly affected by the skill level of the personnel, and it is easy to make mistakes, and the labor cost is high; (4) No self-correction capability of the model: the existing few gas lift control schemes based on fixed models cannot dynamically correct the model parameters according to the changes in the gas well production status. After long-term use, the calculation accuracy decreases and the control effect deteriorates. Summary of the Invention

[0004] To address the aforementioned shortcomings of existing technologies, this invention provides an automatic gas lift control method for shale gas wells, which solves the problems of reliance on manual labor, low control accuracy, and delayed response in existing technologies.

[0005] To achieve the aforementioned objectives, the technical solution adopted by this invention is: an automatic gas lift control method for liquid accumulation in shale gas wells, comprising: S1: Real-time acquisition of shale gas well production parameter data based on wellhead sensors; S2: Based on production parameter data, the liquid accumulation identification logic is used to determine the status of the gas well and obtain the real-time liquid accumulation status. S3: When the real-time liquid accumulation status is determined to be either a liquid accumulation risk state or a liquid accumulation state, the effective production data under the liquid-free state is retrieved from the liquid-free data pool, and the comprehensive critical liquid carrying flow rate is obtained by using a multi-model fusion algorithm. S4: Based on the comprehensive critical liquid carrying flow rate, the optimal gas injection rate is obtained by using a multi-objective optimization algorithm with the goal of maximizing gas injection efficiency; S5: Generate valve opening control commands based on the optimal gas injection volume, use the control commands to drive the gas injection valve to obtain gas lift control results, and complete the gas lift control of shale gas wells.

[0006] Further, S2 includes: S210: Based on the production parameter data of the stable production stage of gas wells without liquid accumulation, the benchmark value for no liquid accumulation and the judgment threshold are calibrated to obtain the judgment threshold. S220: Based on real-time collected production parameter data, the rate of change of each parameter, the oil-casing pressure difference, and the current bottom hole flowing pressure are calculated; S230: The real-time collected production parameter data, parameter change rate, oil casing pressure difference, and current bottom hole flowing pressure are jointly compared with the no-liquidity benchmark value and judgment threshold to determine the real-time liquid accumulation status of the gas well; the real-time liquid accumulation status includes no liquid accumulation status, liquid accumulation risk status, and liquid accumulation status.

[0007] Further, S230 includes: When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference are within the first preset fluctuation range of the no-liquidity benchmark value, and the absolute value of the rate of change of each parameter is less than or equal to the first preset rate of change threshold, the gas well is determined to be in a no-liquidity state. When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference meet the liquid accumulation risk threshold condition in the judgment threshold, and the rate of change of any parameter meets the second preset rate of change condition, the gas well is determined to be in a liquid accumulation risk state. When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference meet the liquid accumulation threshold condition in the judgment threshold, and the duration of this condition is greater than the preset time threshold, the gas well is determined to be in a liquid accumulation state.

[0008] Further, S3 includes: When the situation is determined to be either a liquid accumulation risk state or a liquid accumulation state, the sliding time window combined with the data consistency verification algorithm is used to retrieve the most recent valid production data under the liquid-free state from the liquid-free data pool. The effective production data is input in parallel into multiple critical liquid carrying models to obtain a set of candidate critical flow rates; Based on the dynamic weight calculation strategy, corresponding parameter weights are assigned to each critical liquid-carrying model according to the current operating condition parameters. The comprehensive critical liquid-carrying flow rate is obtained by weighting and fusion calculation of the set of candidate critical flow rates using parameter weights.

[0009] Furthermore, the expression for the parameter weights is: ; ; in, This represents the weight assigned to the i-th critical liquid-carrying model. This represents the fit score of the i-th critical liquid-carrying model calculated based on the current operating parameters. This represents the reliability of the i-th critical liquid-carrying model. This represents the fit score of the j-th critical liquid-carrying model. This represents the reliability of the j-th critical liquid-carrying model. This represents the exponential moving average of the error of the i-th critical liquid-carrying model estimated based on historical liquid accumulation event states. It represents the standard deviation of the error of the i-th critical liquid-carrying model obtained based on the estimation of the state of historical liquid accumulation events.

[0010] Further, S4 includes: Based on effective production data under no liquid accumulation state, the pressure gradient and critical liquid carrying capacity distribution along the wellbore are calculated in segments using a vertical wellbore two-phase flow pressure gradient calculation model, and the bottleneck segment with the minimum liquid carrying margin is identified. Based on the bottleneck section with the minimum liquid carrying margin, and with the constraints of overcoming the liquid carrying demand of the bottleneck section and keeping the bottom hole flowing pressure stable within a reasonable range, an objective function is constructed with the goal of maximizing gas injection efficiency. Based on the comprehensive critical liquid carrying flow rate and the objective function, an iterative search algorithm is used to solve the problem, and the optimal gas injection rate that minimizes the objective function is obtained under the premise of satisfying the constraints.

[0011] Furthermore, the expression for the objective function is: ; in, This represents the value of the unconstrained objective function, including the penalty term. This represents a given candidate gas injection volume. This represents the formation gas supply calculated based on the production capacity response model. Indicates the first penalty coefficient. This represents the function that takes the maximum value. This indicates the preset liquid carrying safety factor. This indicates the liquid carrying capacity of the bottleneck section. Indicates the second penalty coefficient. This indicates the lower limit of the reasonable range for bottom hole flowing pressure. This represents the current bottom hole flowing pressure obtained through iterative calculation. Indicates the third penalty coefficient. This indicates the upper limit of the reasonable range for bottom hole flowing pressure.

[0012] The beneficial effects of this invention are as follows: This invention provides an automatic gas lift control method for shale gas wells with accumulated liquid. By collecting production parameter data in real time based on wellhead sensors and using liquid accumulation identification logic for state determination, it achieves real-time identification and automatic determination of the liquid accumulation state of the gas well. This overcomes the response lag and subjective judgment errors caused by relying on manual inspection in existing technologies, and realizes automatic and accurate triggering of gas lift timing. When it is determined that there is a risk of liquid accumulation or that liquid accumulation has already occurred, effective production data under liquid-free conditions is retrieved from the liquid-free data pool, and a multi-model fusion algorithm is used to calculate the comprehensive critical liquid-carrying flow rate. This effectively avoids the interference of real-time data distortion on the calculation results when the gas well is in abnormal operating conditions. At the same time, the multi-model fusion algorithm overcomes the limitations of a single critical liquid-carrying model under complex operating conditions, significantly improving the accuracy and generalization ability of the critical liquid-carrying flow rate calculation. Based on the comprehensive critical liquid-carrying flow rate, the optimal gas injection rate is calculated using a multi-objective optimization algorithm aimed at maximizing gas injection efficiency, and the control loop is completed after generating control commands to drive the gas injection valve. This process, while ensuring that the bottom-hole fluid carrying bottleneck is effectively overcome and the fluid discharge requirements are met, minimizes the amount of gas injected, significantly reduces the energy consumption of gas lift operations, and improves the overall economic benefits of the gas lift process. Attached Figure Description

[0013] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein: Figure 1 This is an exemplary flowchart of an automatic gas lift control method for liquid accumulation in shale gas wells, as shown in some embodiments of this specification. Detailed Implementation

[0014] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0015] Example Figure 1 This is an exemplary flowchart illustrating an automatic gas lift control method for liquid accumulation in shale gas wells, according to some embodiments of this specification. Figure 1 As shown, the process includes the following steps. In some embodiments, the process may be executed by a processor.

[0016] S1: Real-time acquisition of shale gas well production parameter data based on wellhead sensors.

[0017] Wellhead sensors are hardware devices deployed at the wellhead of shale gas wells to detect physical quantities. For example, wellhead sensors can include pressure sensors, flow sensors, etc.

[0018] In some embodiments, the processor acquires production parameter data of the gas well through the wellhead sensor deployed at the wellhead.

[0019] Production parameter data refers to a set of numerical values ​​that characterize the current or historical operating conditions of shale gas wells. For example, production parameter data can include at least core parameters such as wellhead casing pressure, wellhead oil pressure, and gas production.

[0020] In some embodiments, the processor acquires real-time production parameter data from the wellhead sensor at a preset frequency (e.g., not less than once per minute) through the data acquisition module; the data acquired during the continuous and stable production stage in the early stage of gas well production is used as the production parameter data for the stable production stage of gas well without liquid accumulation.

[0021] S2: Based on production parameter data, the liquid accumulation identification logic is used to determine the status and obtain the real-time liquid accumulation status of the gas well.

[0022] Liquid accumulation detection logic refers to a set of algorithms or judgment rules used to determine whether liquid accumulation has occurred at the bottom of a gas well. For example, liquid accumulation detection logic may include comparison rules for parameter thresholds, evaluation rules for parameter change rates, etc.

[0023] In some embodiments, the processor acquires and executes the liquid accumulation identification logic by reading preset program instructions from the central control module. Production data from the initial 72-hour production phase of the gas well (stable production stage without liquid accumulation) is collected, and the arithmetic mean is taken as the "liquid accumulation baseline value," including baseline casing pressure (P_c0), baseline oil pressure (P_t0), baseline gas production (Q_g0), and baseline oil-casing pressure difference (ΔP0=P_c0-P_t0). The baseline bottom-hole flowing pressure (p_wf0) is obtained through a bottom-hole flowing pressure calculation model. Based on field test data and mature industry experience, two levels of judgment thresholds are calibrated (which can be flexibly adjusted according to the geological conditions and production scale of different gas wells): Liquid accumulation risk threshold: casing pressure P_c≥P_c0×1.15, oil pressure P_t The following conditions must be met: ≤P_t0×0.85, oil-casing pressure difference ΔP≥ΔP0×1.3, gas production Q_g≤Q_g0×0.8, bottom hole flowing pressure p_wf≤p_wf0×0.9 or bottom hole flowing pressure change rate |Δp_wf / Δt|≥1% / min; Any two or more of the above conditions must be met. The following conditions must be met for the liquid accumulation threshold: casing pressure P_c≥P_c0×1.3, oil pressure P_t≤P_t0×0.7, oil-casing pressure difference ΔP≥ΔP0×1.5, gas production Q_g≤Q_g0×0.7, bottom hole flowing pressure p_wf≤p_wf0×0.8 or bottom hole flowing pressure continuously decreasing for more than 10 minutes; Any two or more of the above conditions must be met.

[0024] Real-time liquid accumulation status refers to the liquid-carrying capacity level of a gas well at the current moment. For example, real-time liquid accumulation status can include no liquid accumulation status (indicating that the gas well is carrying liquid normally), liquid accumulation risk status (indicating that the liquid-carrying capacity is declining and there is a tendency for liquid accumulation), and liquid accumulation status (indicating that liquid accumulation has formed at the bottom of the well, affecting production).

[0025] In some embodiments, the processor calculates and determines the real-time liquid accumulation status by inputting real-time collected production parameter data into the liquid accumulation identification logic.

[0026] In some embodiments, the processor may implement S2 based on the following steps.

[0027] S210: Based on the production parameter data of the stable production stage of gas wells without liquid accumulation, the benchmark value for no liquid accumulation and the judgment threshold are calibrated.

[0028] The fluid-free benchmark refers to the reference standard for various parameters of a gas well in a healthy state where there is absolutely no fluid accumulation. For example, the fluid-free benchmark may include the arithmetic mean of data such as benchmark casing pressure, benchmark oil pressure, benchmark gas production, benchmark oil-casing pressure difference, and benchmark bottom hole flowing pressure.

[0029] In some embodiments, the processor obtains the no-liquidity baseline value by performing an arithmetic average of the production parameter data from the stable production phase of the gas well during the initial 72-hour period without liquid accumulation.

[0030] The judgment threshold refers to the set of critical values ​​used to trigger state switching boundaries. For example, the judgment threshold may include the values ​​corresponding to the liquid accumulation risk threshold condition and the values ​​corresponding to the liquid accumulation threshold condition, such as the percentage increase in casing pressure and the percentage decrease in oil pressure.

[0031] In some embodiments, the processor calculates and obtains the judgment threshold based on the no-liquidity benchmark value and field test data according to a preset ratio.

[0032] S220: Based on real-time collected production parameter data, the rate of change of each parameter, the oil-casing pressure difference, and the current bottom hole flowing pressure are calculated.

[0033] The rate of change of a parameter refers to the magnitude of change of production parameter data per unit time. For example, the rate of change of a parameter may include the rate of change of casing pressure, the rate of change of oil pressure, the rate of change of gas production, and the rate of change of bottom hole flowing pressure (e.g., % / min).

[0034] In some embodiments, the processor obtains the rate of change of the production parameter by performing differentiation or difference calculation on the production parameter data at adjacent time points or within a sliding time window.

[0035] The casing pressure difference refers to the difference between the casing pressure and the tubing pressure in a gas well.

[0036] In some embodiments, the processor calculates the oil-casing pressure difference by subtracting the wellhead oil pressure from the real-time collected wellhead casing pressure.

[0037] Bottom-hole flowing pressure refers to the pressure exerted on the well wall by gas flowing at the bottom of the well.

[0038] In some embodiments, the processor calculates the current bottom hole flowing pressure based on real-time acquired wellhead casing pressure, wellhead oil pressure, and gas well physical structure data, using a preset bottom hole flowing pressure calculation model. It receives raw parameter data transmitted from the data acquisition module, uses a moving average filtering algorithm to filter out instantaneous interference signals, and extracts valid data; it calculates the oil-casing pressure difference (ΔP=P_c-P_t) and the rate of change of each parameter, and simultaneously calls the bottom hole flowing pressure calculation module to calculate the current bottom hole flowing pressure (p_wf) in real time.

[0039] S230: The real-time collected production parameter data, parameter change rate, oil casing pressure difference, and current bottom hole flowing pressure are jointly compared with the no-liquidity benchmark value and judgment threshold to determine the real-time liquid accumulation status of the gas well; the real-time liquid accumulation status includes no liquid accumulation status, liquid accumulation risk status, and liquid accumulation status.

[0040] In some embodiments, the processor may make a judgment based on the following logic: when the real-time acquired production parameter data, the current bottom hole flowing pressure, and the oil-casing pressure difference are within a first preset fluctuation range of the no-liquidity benchmark value, and the absolute value of the rate of change of each parameter is less than or equal to a first preset rate of change threshold, the gas well is determined to be in a no-liquidity state; when the real-time acquired production parameter data, the current bottom hole flowing pressure, and the oil-casing pressure difference meet the liquid accumulation risk threshold condition in the judgment threshold, and the rate of change of any parameter meets a second preset rate of change condition, the gas well is determined to be in a liquid accumulation risk state; when the real-time acquired production parameter data, the current bottom hole flowing pressure, and the oil-casing pressure difference meet the liquid accumulation threshold condition in the judgment threshold, and the duration of this satisfied state is greater than a preset time threshold, the gas well is determined to be in a liquid accumulation state.

[0041] The first preset fluctuation range refers to the reasonable range within which gas well parameters are allowed to fluctuate normally under conditions of no fluid accumulation. For example, the first preset fluctuation range can be ±10% of the baseline value without fluid accumulation, or ±5% of the baseline bottomhole flowing pressure.

[0042] In some embodiments, the processor obtains the first preset fluctuation range by reading preset values ​​pre-stored in the system.

[0043] The first preset rate of change threshold refers to the upper limit of the parameter change rate that characterizes the gas well operating conditions in a stable state. For example, the first preset rate of change threshold may include the rate of change of conventional parameters with an absolute value ≤0.5% / min and the rate of change of bottom hole flowing pressure ≤0.3% / min.

[0044] In some embodiments, the processor obtains the first preset rate of change threshold by reading system configuration parameters.

[0045] The fluid accumulation risk threshold condition refers to the deterioration boundary of parameters that characterize the impending fluid accumulation in a gas well. For example, this condition may include casing pressure greater than or equal to 1.15 times the benchmark value, oil pressure less than or equal to 0.85 times the benchmark value, etc.

[0046] In some embodiments, the processor obtains the fluid accumulation risk threshold condition by reading a pre-calibrated fluid accumulation risk threshold set by the system.

[0047] The second preset rate of change condition refers to the rate of change that characterizes the rapid deterioration of gas well parameters. For example, the second preset rate of change condition may include casing pressure change rate ≥ 0.3% / min, oil pressure change rate ≤ -0.3% / min, etc.

[0048] In some embodiments, the processor obtains the second preset rate of change condition by reading system configuration parameters.

[0049] The liquid accumulation threshold condition refers to the parameter deterioration boundary that characterizes a gas well where severe liquid accumulation has occurred. For example, this condition may include casing pressure greater than or equal to 1.3 times the benchmark value, oil pressure less than or equal to 0.7 times the benchmark value, etc.

[0050] In some embodiments, the processor obtains the accumulated fluid threshold condition by reading the accumulated fluid threshold pre-calibrated by the system.

[0051] A time threshold refers to the duration required for a state to be considered to have occurred. For example, a time threshold could be 5 minutes.

[0052] In some embodiments, the processor obtains the time threshold by counting the time the state is maintained through a built-in clock module.

[0053] In some embodiments, the processor compares the preprocessed real-time parameters, the calculated oil-casing pressure difference, the bottom hole flowing pressure, and the rate of change of each parameter with the calibrated thresholds to achieve accurate classification of the gas well state: No fluid accumulation state: Casing pressure P_c∈P_c0±10%, Oil pressure P_t∈P_t0±10%, ΔP∈ΔP0±10%, Q_g∈Q_g0±10%, p_wf∈p_wf0±5%, and the absolute value of the rate of change of each parameter ≤0.5% / min (bottom hole flowing pressure) Change rate ≤ 0.3% / min); Liquid accumulation risk status: meets any 2 or more of the "Liquid accumulation risk thresholds", and the core change rate (ΔP_c / Δt ≥ 0.3% / min, ΔP_t / Δt ≤ -0.3% / min, ΔQ_g / Δt ≤ -0.5% / min, Δp_wf / Δt ≤ -0.5% / min) meets any 1 of the following; Liquid accumulation status: meets any 2 or more of the "Liquid accumulation thresholds", and the duration of this status is ≥ 5 minutes.

[0054] S3: When the real-time liquid accumulation status is determined to be either a liquid accumulation risk state or a liquid accumulation state, the effective production data under the liquid-free state is retrieved based on the liquid-free data pool, and the comprehensive critical liquid carrying flow rate is obtained by using a multi-model fusion algorithm.

[0055] A liquid-free data pool refers to a database or storage area that stores various parameters of gas wells under historical healthy production conditions and their derived calculation results. For example, a liquid-free data pool can include a large amount of valid production data from historical liquid-free conditions.

[0056] In some embodiments, the processor accumulates and saves production parameter data determined to be in a liquid-free state through a historical data storage unit to construct the liquid-free data pool. After the gas injection operation, when the real-time parameters recover to the "liquid-free baseline value ±10%" range and remain stable for ≥10 minutes, it is determined that the liquid has been completely discharged; for every 1000 sets of stable production data in a liquid-free state, the "liquid-free baseline value" (including p_wf0) is automatically updated, and the threshold is adjusted synchronously according to the original proportional relationship; at the same time, the historical data storage unit stores the effective production data (including parameters, calculation models and results) in a liquid-free state in real time, forming a "liquid-free data pool" for subsequent use.

[0057] Effective production data refers to representative, stable operating condition data extracted from the fluid-free data pool that has not been affected by transient disturbances. For example, effective production data may include casing pressure, oil pressure, gas production, well inclination angle, etc., for a certain historical period.

[0058] In some embodiments, the processor obtains the valid production data by performing a sliding time window combined with a data consistency verification algorithm on the data in the data pool.

[0059] Multi-model fusion algorithms refer to computational strategies or program modules used to integrate the results of multiple individual prediction models. For example, multi-model fusion algorithms may include adaptive weight allocation strategies, multi-model parallel computing architectures, etc.

[0060] In some embodiments, the processor obtains and executes the multi-model fusion algorithm by invoking algorithm instructions preset in the central control module.

[0061] The comprehensive critical fluid-carrying capacity refers to the minimum necessary gas production rate required to ensure effective discharge of accumulated fluid from a gas well under current operating conditions. For example, the comprehensive critical fluid-carrying capacity can be a weighted sum of candidate critical flow rate values ​​calculated by multiple individual models.

[0062] In some embodiments, the processor calculates the comprehensive critical liquid-carrying flow rate by inputting the effective production data into the multi-model fusion algorithm. The final comprehensive critical liquid-carrying flow rate q_cr_fused is obtained by weighted summation: q_cr_fused = Σ(ω_i * q_cri). This method effectively utilizes the advantages of each model under different operating conditions, improving the accuracy and adaptability of the calculation.

[0063] In some embodiments, when the condition is determined to be either a liquid accumulation risk state or a liquid accumulation state, the processor can use a sliding time window combined with a data consistency verification algorithm to retrieve the most recent valid production data under the liquid-free state from the liquid-free data pool; input the valid production data in parallel into multiple critical liquid-carrying models to obtain a set of candidate critical flow values; based on a dynamic weight calculation strategy, assign corresponding parameter weights to each critical liquid-carrying model according to the current operating parameters; and use the parameter weights to perform weighted fusion calculation on the set of candidate critical flow values ​​to obtain the comprehensive critical liquid-carrying flow rate.

[0064] Critical fluid carrying capacity models are mathematical or physical models used to predict the fluid carrying capacity of gas wells. Examples of critical fluid carrying capacity models include, but are not limited to, the Turner, Coleman, LiMin, and Peng Zhaoyang models.

[0065] In some embodiments, the processor obtains the critical liquid-carrying model by invoking a pre-configured mathematical or empirical formula.

[0066] Candidate critical flow rates refer to predicted values ​​calculated independently by a single critical fluid-carrying model. For example, a candidate critical flow rate can be a flow rate value calculated by a specific model based on the current operating conditions. The retrieved no-fluid-accumulation operating condition parameter vector G=[bottom hole pressure, gas-liquid ratio, well inclination angle,...] is input in parallel into multiple critical fluid-carrying models (including but not limited to Turner, Coleman, LiMin, and PengZhaoYang models) to obtain a set of candidate critical flow rates Q_cr_set.

[0067] In some embodiments, the processor obtains a set of candidate critical flow values ​​by inputting the effective production data into different critical liquid-carrying models for parallel computation.

[0068] Dynamic weighting calculation strategy refers to the algorithmic rules that assign reliability scores to different models based on the current operating characteristics of the gas well. For example, dynamic weighting calculation strategy may include smoothing operating condition factors constructed based on operating condition parameters such as bottom hole flowing pressure, gas-liquid ratio, and well inclination angle, as well as calculation formulas for reliability based on historical errors.

[0069] In some embodiments, the processor obtains the dynamic weight calculation strategy by calling a weight calculation program preset in the gas injection parameter calculation unit. A dynamic weight calculation strategy is designed to assign appropriate weights ω_i to each model based on the current operating condition G. (Example of weight strategy: In the high-pressure range, increase the weights of the Turner and PengZhaoYang models; in the high gas-liquid ratio range, increase the weights of the Coleman model; for highly deviated wells, significantly increase the weights of the LiMin model.) The dynamic weight fusion strategy includes: constructing a smoothing operating condition factor based on operating condition parameters such as bottomhole flowing pressure p_wf, gas-liquid ratio GLR, and well inclination angle θ. And calculate the fit score for each critical liquid-carrying model i. Furthermore, the system estimates the critical boundary of observation based on the state switching of historical effusion events. Calculate the exponential moving average of the errors of each model. To obtain reliability The final weights are calculated as follows: The calculation enables adaptive allocation and online self-correction of model contributions under different operating conditions.

[0070] Parameter weights are coefficient values ​​that characterize the contribution or importance of an individual model in fusion computation. For example, a parameter weight can be a value between 0 and 1 assigned to a model based on specific operating conditions such as high pressure or steep incline.

[0071] In some embodiments, the processor calculates the parameter weights corresponding to each critical liquid-carrying model for the current operating condition parameters by executing the dynamic weight calculation strategy.

[0072] In some embodiments, the expression for the parameter weights is: ; ; in, This represents the weight assigned to the i-th critical liquid-carrying model. This represents the fit score of the i-th critical liquid-carrying model calculated based on the current operating parameters. This represents the reliability of the i-th critical liquid-carrying model. This represents the fit score of the j-th critical liquid-carrying model. This represents the reliability of the j-th critical liquid-carrying model. This represents the exponential moving average of the error of the i-th critical liquid-carrying model estimated based on historical liquid accumulation event states. It represents the standard deviation of the error of the i-th critical liquid-carrying model obtained based on the estimation of the state of historical liquid accumulation events.

[0073] In some embodiments, the processor can trigger a model parameter correction mechanism when the accumulated effective production data under the fluid-free state reaches a set quantity or period: a monitoring signal is constructed based on the effective production data under the fluid-free state, and the weight parameters in the critical fluid-carrying model and the key parameters in the vertical wellbore two-phase flow pressure gradient calculation model are dynamically updated using the monitoring signal. Only stable and effective data in the "fluid-free data pool" are used as the basis for model correction. The model correction unit continuously extracts new fluid-free data and adjusts the friction coefficient of the vertical wellbore two-phase flow pressure gradient model accordingly. (or equivalent correction coefficient for friction term) and weight parameters in the critical liquid carrying capacity fusion model Key parameters are updated regularly, and a deviation threshold trigger mechanism is set: when the model predicts the comprehensive critical liquid carrying capacity... The actual liquid carrying capacity calculated from the state without liquid accumulation The relative deviation exceeds When needed, parameters are automatically adjusted to ensure that the model closely reflects real-world working conditions over the long term.

[0074] 1) Construction of a fluid-free data pool and monitoring signals. Within the fluid-free stable section, the wellbore can be considered to meet the constraint of "capable of carrying fluid without continuous fluid accumulation." Therefore, two types of monitoring signals can be constructed for correction: Pressure distribution monitoring: Bottomhole flowing pressure calculated from the wellhead casing pressure / oil pressure, wellbore structure, and PVT (Pressure-Volume-Temperature) properties using a segmented two-phase flow model. Inversion / verification based on a fluid-free baseline Error; Liquid carrying capacity monitoring: Under the condition of no liquid accumulation, the actual total rising gas volume can be approximately expressed as... And it should meet the liquid carrying capacity requirement. (For safety factors). Based on this, the "critical boundary of the liquid-free state" is defined as... As a monitoring signal for the overall critical liquid carrying capacity.

[0075] 2) Least squares correction path: Introduce a correctable coefficient into the friction term in the pressure gradient model.

[0076] .

[0077] In a sample window without fluid accumulation Internally, construct the least squares objective: ; Seeking Later update (or update the equivalent coefficient of the friction term), and similarly introduce low-dimensional interpretable corrections to the weight parameters (such as adding a bias to the scoring term before softmax). And perform least squares / ridge regression fitting to achieve fast and robust online drift correction.

[0078] in, It is the total pressure gradient per unit depth within the wellbore, expressed in MPa / m. It is the pressure gradient term (gravity term) generated by fluid gravity. It is the pressure gradient term (friction term) generated by the friction between the fluid and the wellbore wall. It is the pressure gradient term (acceleration term) generated by the fluid acceleration effect; It is the friction term correction coefficient, a dimensionless parameter used to correct the wellbore friction pressure gradient online, in order to compensate for the effects of wellbore roughness changes, flow regime changes or model errors; The non-liquid sample window represents a set of continuous time samples selected from the non-liquid data pool; At the friction correction coefficient of At that time, the predicted bottom hole flowing pressure was calculated using the wellbore two-phase flow pressure gradient model; These are bottom hole flowing pressure observations obtained through production data inversion or on-site measurements.

[0079] 3) Neural Network Correction Path: When wellbore conditions are complex and parameter drift is nonlinear, this invention employs a lightweight multi-task neural network to jointly learn key correction quantities. The network structure uses a multilayer perceptron (MLP) with multiple outputs, and the input is a feature vector of the fluid-free operating condition. ; Network output includes: (1) Friction correction factor:

[0080] (2) Fusion weight vector: .

[0081] The network training employs a "weighted summation of multi-objective loss functions," where each constraint objective corresponds to a loss term: the voltage stabilization objective loss (pressure prediction consistency). ; Liquid carrying target loss (critical liquid carrying flow rate close to the back-calculated value without liquid accumulation) ; Discharge safety constraint loss (liquid carrying margin must meet the threshold) ; Parameter smoothing loss (suppressing frequent and drastic adjustments) ; Regularization loss (to prevent overfitting) ; in These are network parameters.

[0082] The final total loss function is: ; in, These are configurable weighting coefficients used to balance pressure stability consistency, liquid carrying accuracy, safety margin, parameter smoothness, and model generalization ability. Bottom hole flowing pressure; For casing pressure; For the oil pipe pressure; This is the oil-casing pressure differential, which is the difference between the oil tubing pressure and the casing pressure. For gas production; The gas-liquid ratio; The well inclination angle; This refers to the inner diameter of the oil pipe. The temperature of the fluid in the wellbore; The density of the liquid phase; This refers to the dynamic viscosity of the liquid phase. This refers to the gas phase density. This refers to the dynamic viscosity of the gas phase. The coefficient for voltage regulation consistency loss is used. The pressure stabilization target loss is used to measure the deviation between the predicted bottom-hole flowing pressure and the observed bottom-hole flowing pressure. The critical liquid carrying flow rate loss is used to measure the relative error between the predicted critical liquid carrying flow rate and the back-calculated critical liquid carrying flow rate. This is a safety constraint loss for fluid discharge, used to ensure that the fluid carrying capacity of the gas well meets the safety margin constraint; This is a parameter smoothing loss used to limit drastic changes in weights and friction correction factors between adjacent time steps; This is the regularization loss, used to suppress model complexity and prevent overfitting.

[0083] 4) Triggering and update strategy, (1) Periodic update: every cumulative (2) Threshold trigger: If the condition is met within the nearest window, the model parameters are updated once a new sample without effusion is added to the group (or once every fixed period such as 24h);

[0084] This will trigger a fix immediately and update it first. and Once the correction is complete, the updated parameters will be written to the historical version repository for rollback and auditing.

[0085] in, This is the comprehensive critical liquid carrying flow rate prediction obtained by integrating multiple critical liquid carrying models; The observed critical liquid-carrying flow rate, calculated from the stable operating condition without liquid accumulation, is used as a supervisory reference value for model calibration. This is a numerically stable term (a very small positive number), used to avoid calculation instability caused by a denominator of zero or a value that is too small; It is a friction correction factor used to correct the friction term in the two-phase flow pressure gradient model of the wellbore, in order to compensate for deviations caused by changes in wellbore roughness, flow regime changes, or model structure errors. This is the fusion weight vector for the critical liquid-carrying model.

[0086] S4: Based on the comprehensive critical liquid carrying flow rate, the optimal gas injection rate is obtained by using a multi-objective optimization algorithm with the goal of maximizing gas injection efficiency.

[0087] Multi-objective optimization algorithms are computational methods used to find the optimal compromise solution among multiple conflicting objectives. For example, multi-objective optimization algorithms may include iterative search programs that focus on maximizing gas injection efficiency while incorporating constraints for liquid drainage and pressure stabilization.

[0088] In some embodiments, the processor obtains and executes the multi-objective optimization algorithm by invoking an optimization program preset in the central control module.

[0089] The optimal gas injection rate refers to the gas injection volume or mass flow rate that maximizes injection efficiency and minimizes energy consumption while satisfying the constraints of fluid carrying capacity and pressure stability in a gas well. For example, the optimal gas injection rate could be a candidate gas injection rate that minimizes the objective function value, including the penalty term.

[0090] In some embodiments, the processor calculates the optimal gas injection rate based on the comprehensive critical liquid carrying flow rate and the objective function using an iterative search algorithm.

[0091] In some embodiments, the processor can use effective production data under a fluid-free state to calculate the pressure gradient and critical fluid-carrying capacity distribution along the wellbore in segments using a vertical wellbore two-phase flow pressure gradient calculation model, and identify the bottleneck segment with the minimum fluid-carrying margin. Based on the bottleneck segment with the minimum fluid-carrying margin, and with the constraints of overcoming the fluid-carrying demand of the bottleneck segment and keeping the bottom hole pressure stable within a reasonable range, an objective function is constructed with the goal of maximizing gas injection efficiency. Based on the comprehensive critical fluid-carrying flow rate and the objective function, an iterative search algorithm is used to solve the problem, and the optimal gas injection rate that minimizes the objective function is obtained under the premise of satisfying the constraints.

[0092] A pressure gradient calculation model for two-phase flow in a vertical wellbore is a physical model used to simulate the pressure variation along the flow path when gas and liquid phases flow vertically within a wellbore. For example, this model may include iterative formulas for calculating the liquid holdup, friction coefficient, and miscibility density within a section.

[0093] In some embodiments, the processor obtains the pressure gradient calculation model of the two-phase flow in the vertical wellbore by invoking a pre-configured fluid dynamics model instruction.

[0094] The bottleneck section with the smallest fluid carrying capacity refers to the section in the gas wellbore where the ratio of the actual gas rising capacity to the required critical fluid carrying capacity is the smallest; that is, the weakest link most prone to fluid accumulation. For example, the bottleneck section can be the section with the calculated minimum fluid carrying capacity among the discrete sections distributed along the wellbore.

[0095] In some embodiments, the processor, based on the effective production data, calculates the bottleneck segment with the smallest fluid carrying margin by segmenting along the wellbore using the vertical wellbore two-phase flow pressure gradient calculation model and comparing the fluid carrying margin.

[0096] Constraints refer to physical or operational boundary limitations that must be met during optimization calculations. For example, constraints may include drainage constraints to overcome the fluid carrying requirements of the bottleneck section (fluid carrying margin reaching the safety factor) and pressure stabilization constraints to ensure that the bottom hole flowing pressure remains stable within a reasonable range.

[0097] In some embodiments, the processor obtains the constraints by reading preset engineering standards and safety boundaries from the system. Liquid discharge (liquid carrying) constraint: The liquid carrying margin at the bottleneck section must meet the safety factor. Bottom-hole flowing pressure stabilization constraint: The bottom-hole flowing pressure is maintained within a reasonable range. Execution and safety constraints: Injection volume and valve actuation are limited. .

[0098] An objective function is a mathematical expression that needs to be minimized or maximized during the optimization process. For example, an objective function can be an unconstrained function that minimizes the proportion of gas injection (maximizes gas injection efficiency) and incorporates a penalty term for violating constraints (as described in the formula in claim 9).

[0099] In some embodiments, the processor obtains the target function by invoking a preset mathematical expression.

[0100] In some embodiments, the expression for the objective function is: ; in, This represents the value of the unconstrained objective function, including the penalty term. This represents a given candidate gas injection volume. This represents the formation gas supply calculated based on the production capacity response model. Indicates the first penalty coefficient. This represents the function that takes the maximum value. This indicates the preset liquid carrying safety factor. This indicates the liquid carrying capacity of the bottleneck section. Indicates the second penalty coefficient. This indicates the lower limit of the reasonable range for bottom hole flowing pressure. This represents the current bottom hole flowing pressure obtained through iterative calculation. Indicates the third penalty coefficient. This indicates the upper limit of the reasonable range for bottom hole flowing pressure.

[0101] Iterative search algorithms are numerical computation methods that approximate the optimal solution by repeatedly substituting candidate solutions. For example, iterative search algorithms can include two-stage methods such as feasible region boundary search (doubling strategy, etc.) and interval minimization search (such as golden section Brent search, etc.).

[0102] In some embodiments, the processor obtains and executes the iterative search algorithm by invoking a search program preset in the gas injection parameter calculation unit. The iterative algorithm solves for the optimal gas injection rate q_g_opt that minimizes the objective function J while satisfying the constraints of liquid drainage and pressure stabilization. This scheme aims to maximize gas injection savings and reduce energy consumption while ensuring stable production. The specific method is as follows: (1) Factors affecting gas injection volume and calculation link (updated in each iteration), each time a candidate gas injection volume is given. (or valve opening) The system calculates and updates key quantities in the following order: Valve / throttling model (gas injection volume generation): if based on valve opening... To control the flow rate, it is converted into the injection flow rate using the injection valve flow model: in For the gas injection source pressure, For hydraulic pressure These are gas state parameters.

[0103] Iterative solution of pressure gradient in segmented two-phase flow in wellbore (obtained) and Discretize the wellbore into Section (Section Length) The pressure distribution is solved iteratively segment by segment from the wellhead downwards. For the first... Segment, at a given segment inlet pressure In the case of iterative calculation of liquid holdup within the segment coefficient of friction Miscibility This leads to a pressure drop. And update the export pressure of the section. Finally, the bottom hole flowing pressure is obtained.

[0104] Calculation of critical fluid carrying capacity distribution along the wellbore and identification of bottleneck sections: For each section, under the pressure, temperature, and fluid properties conditions of that section, the critical fluid carrying flow rate of that section is obtained by calling "multi-model fusion critical fluid carrying flow rate". Define the actual total rising gas volume in this segment. Liquid carrying capacity: The section with the smallest margin in the entire well is taken as the "bottleneck section". .in, The wellbore is segmented and numbered to indicate the k-th calculation segment after the wellbore is divided along the depth direction; This is the set of operating parameters for the k-th well section, including information such as pressure, temperature, gas-liquid ratio, well inclination angle, and fluid properties for that section, used to calculate the model weights and critical fluid carrying capacity for that section. The comprehensive critical fluid carrying capacity of the k-th well section is obtained by weighted fusion of multiple critical fluid carrying models; For the i-th critical liquid-carrying model, the operating conditions in segment k are... The sum of the fusion weights under; The critical fluid carrying flow rate is calculated for the i-th critical fluid carrying model under the k-th wellbore conditions; σ_e represents the liquid carrying margin index of the k-th wellbore, used to measure the safety of the actual gas flow rate relative to the critical liquid carrying flow rate; j represents the traversal index of the critical liquid carrying model, i.e., the j-th model; σ_e represents the standard deviation of the critical liquid carrying model error estimated based on the historical liquid accumulation event state; t represents the time index of the continuous time sample set within the liquid-free sample window; g_f(x) represents the pre-linear mapping function of the output branch corresponding to the friction correction factor in the neural network; g_ω(x) represents the pre-linear mapping function of the output branch corresponding to the fusion weight vector in the neural network; ε represents a very small positive number introduced to avoid computational instability caused by a denominator of zero or too small a value, i.e., a numerical stability term; Δt represents the time difference between adjacent time steps; Θ represents the set of network parameters used to control the output and fitting effect of the multi-task neural network model; production capacity response model ( Mapped to formation gas supply ): Introducing a capacity model: in For production parameters that can be corrected online, "gas injection changes wellbore pressure drop" Change Change The coupling relationship was incorporated into the optimization.

[0105] In some embodiments, since the decision variables are one-dimensional The algorithm employs a two-stage iterative approach: "feasible region search + interval line search optimization". The first stage involves searching the feasible region boundary (determining the feasible interval) and setting initial values. and upper and lower boundaries .like Then gradually increase (e.g., doubling strategy) Until the constraints are met or the upper limit is reached. Thus, the feasible interval is obtained.

[0106] Iterative minimization within the feasible interval (finding the optimal solution): using the golden ratio within the feasible interval. Search minimum The point. Each iteration on the candidate Execute "valve model" Two-phase flow in well section The complete evaluation and narrowing of the interval for the objective and penalty function is performed until the stopping condition is met:

[0107]

[0108]

[0109] Final output This information is then sent to the execution module to achieve automatic gas injection control. The gas flow rate or gas volume control is a core decision variable in the optimization calculation.

[0110] in, The opening degree of the gas injection valve or the adjustment parameter of the control valve has a corresponding relationship with the gas injection flow rate q_g; To optimize the objective function, it is used to comprehensively evaluate the performance of a gas well under a certain gas injection flow rate. This represents the objective function value at the k-th iteration. The gas injection flow rate is calculated in the k-th iteration; k is the current iteration number. This is the maximum allowed number of iterations, used to prevent the optimization algorithm from running for a long time or getting stuck in a loop.

[0111] S5: Generate valve opening control commands based on the optimal gas injection volume, use the control commands to drive the gas injection valve to obtain gas lift control results, and complete the gas lift control of shale gas wells.

[0112] Valve opening control commands refer to electrical signals or digital instructions used to drive electric gas injection valves to perform actions. For example, a valve opening control command may contain a control code that adjusts the valve opening to a target percentage between 0% and 100%.

[0113] In some embodiments, the processor converts the calculated optimal gas injection volume into a target opening degree based on the valve flow model, and then generates a control command to obtain the valve opening degree.

[0114] Gas lift control results refer to the actual changes in the state of liquid accumulation and production parameters in the gas well after the system performs gas injection operations. For example, gas lift control results may include whether the accumulated liquid has been discharged, casing pressure, oil pressure, and gas production have returned to normal ranges.

[0115] In some embodiments, after the processor confirms that the valve opening is in place through the signal feedback module, it continuously monitors the real-time production parameter data of the gas well to obtain the gas lift control result.

[0116] In some embodiments, the actuator consists of an injection pipeline and an electric injection valve. The electric injection valve is deployed at the connection between the injection pipeline and the well casing and is used to receive valve opening instructions issued by the central control module and drive the valve to the target opening. The opening adjustment range of the electric injection valve is 0-100%, and the adjustment accuracy is ±1%, which can adapt to the control requirements of different injection volumes.

[0117] In some embodiments, the signal feedback device is integrated inside the electric gas injection valve. A displacement sensor is used to detect the actual opening degree of the valve in real time. When the deviation between the actual opening degree and the target opening degree is ≤±1% (i.e., the valve is in the correct position), the "opening degree in the correct position" signal is sent back to the central control module via the RS485 bus to ensure that the gas injection process is strictly executed according to the instructions.

Claims

1. A method for automatic gas lift control of liquid accumulation in shale gas wells, characterized in that, include: S1: Real-time acquisition of shale gas well production parameter data based on wellhead sensors; S2: Based on production parameter data, the liquid accumulation identification logic is used to determine the status of the gas well and obtain the real-time liquid accumulation status. S3: When the real-time liquid accumulation status is determined to be either a liquid accumulation risk state or a liquid accumulation state, the effective production data under the liquid-free state is retrieved from the liquid-free data pool, and the comprehensive critical liquid carrying flow rate is obtained by using a multi-model fusion algorithm. S4: Based on the comprehensive critical liquid carrying flow rate, the optimal gas injection rate is obtained by using a multi-objective optimization algorithm with the goal of maximizing gas injection efficiency; S5: Generate valve opening control commands based on the optimal gas injection volume, use the control commands to drive the gas injection valve to obtain gas lift control results, and complete the gas lift control of shale gas wells.

2. The automatic gas lift control method for shale gas well liquid accumulation according to claim 1, characterized in that, S2 includes: S210: Based on the production parameter data of the stable production stage of gas wells without liquid accumulation, the benchmark value for no liquid accumulation and the judgment threshold are calibrated to obtain the judgment threshold. S220: Based on real-time acquired production parameter data, the rate of change of each parameter, the oil-casing pressure difference, and the current bottom hole flowing pressure are calculated; S230: The real-time collected production parameter data, parameter change rate, oil casing pressure difference, and current bottom hole flowing pressure are jointly compared with the no-liquidity benchmark value and judgment threshold to determine the real-time liquid accumulation status of the gas well; the real-time liquid accumulation status includes no liquid accumulation status, liquid accumulation risk status, and liquid accumulation status.

3. The automatic gas lift control method for shale gas well liquid accumulation according to claim 2, characterized in that, S230 includes: When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference are within the first preset fluctuation range of the no-liquidity benchmark value, and the absolute value of the rate of change of each parameter is less than or equal to the first preset rate of change threshold, the gas well is determined to be in a no-liquidity state. When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference meet the liquid accumulation risk threshold condition in the judgment threshold, and the rate of change of any parameter meets the second preset rate of change condition, the gas well is determined to be in a liquid accumulation risk state. When the real-time collected production parameter data, the current bottom hole flowing pressure, and the oil casing pressure difference meet the liquid accumulation threshold condition in the judgment threshold, and the duration of this condition is greater than the preset time threshold, the gas well is determined to be in a liquid accumulation state.

4. The automatic gas lift control method for shale gas well liquid accumulation according to claim 1, characterized in that, S3 includes: When the situation is determined to be either a liquid accumulation risk state or a liquid accumulation state, the sliding time window combined with the data consistency verification algorithm is used to retrieve the most recent valid production data under the liquid-free state from the liquid-free data pool. The effective production data is input in parallel into multiple critical liquid carrying models to obtain a set of candidate critical flow rates; Based on the dynamic weight calculation strategy, corresponding parameter weights are assigned to each critical liquid-carrying model according to the current operating condition parameters. The comprehensive critical liquid-carrying flow rate is obtained by weighting and fusion calculation of the set of candidate critical flow rates using parameter weights.

5. The automatic gas lift control method for shale gas well liquid accumulation according to claim 4, characterized in that, The expression for the parameter weights is: ； ； in, This represents the weight assigned to the i-th critical liquid-carrying model. This represents the fit score of the i-th critical liquid-carrying model calculated based on the current operating parameters. This represents the reliability of the i-th critical liquid-carrying model. This represents the fit score of the j-th critical liquid-carrying model. This represents the reliability of the j-th critical liquid-carrying model. This represents the exponential moving average of the error of the i-th critical liquid-carrying model estimated based on historical liquid accumulation event states. It represents the standard deviation of the error of the i-th critical liquid-carrying model obtained based on the estimation of the state of historical liquid accumulation events.

6. The automatic gas lift control method for shale gas well liquid accumulation according to claim 1, characterized in that, S4 includes: Based on effective production data under no liquid accumulation state, the pressure gradient and critical liquid carrying capacity distribution along the wellbore are calculated in segments using a vertical wellbore two-phase flow pressure gradient calculation model, and the bottleneck segment with the minimum liquid carrying margin is identified. Based on the bottleneck section with the minimum liquid carrying margin, and with the constraints of overcoming the liquid carrying demand of the bottleneck section and keeping the bottom hole flowing pressure stable within a reasonable range, an objective function is constructed with the goal of maximizing gas injection efficiency. Based on the comprehensive critical liquid carrying flow rate and the objective function, an iterative search algorithm is used to solve the problem, and the optimal gas injection rate that minimizes the objective function is obtained under the premise of satisfying the constraints.

7. The automatic gas lift control method for shale gas well liquid accumulation according to claim 6, characterized in that, The expression for the objective function is: ； in, This represents the value of the unconstrained objective function, including the penalty term. This represents a given candidate gas injection volume. This represents the formation gas supply calculated based on the production capacity response model. Indicates the first penalty coefficient. This represents the function that takes the maximum value. This indicates the preset liquid carrying safety factor. This indicates the liquid carrying capacity of the bottleneck section. Indicates the second penalty coefficient. This indicates the lower limit of the reasonable range for bottom hole flowing pressure. This represents the current bottom hole flowing pressure obtained through iterative calculation. Indicates the third penalty coefficient. This indicates the upper limit of the reasonable range for bottom hole flowing pressure.

Images