Platform circulation gas lift intelligent optimization and edge control cooperation method

Abstract

The invention discloses a platform circulation gas lift intelligent optimization and edge control cooperation method. The method comprises the steps of 1, drawing a single well gas lift dynamic curve, 2, establishing a multi-objective optimization model, 3, optimizing and dynamically adjusting gas injection parameters, and 4, carrying out circulation gas lift intelligent control. According to the method, an integrated technical system of'single well dynamic modeling-multi-well collaborative optimization-edge intelligent control 'is constructed, global optimal distribution of the gas injection amount is achieved through a two-stage optimization framework and a particle swarm optimization algorithm, the problem of coupling interference existing in a multi-well platform is solved, real-time response is achieved through edge control, and the method is suitable for large-scale popularization and application. According to the method, the targets of accurate control and yield maximization of the circulating gas lift process can be achieved, and a reliable technical path is provided for intelligent drainage gas recovery of the shale gas field.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of oil and gas exploitation, and particularly relates to a platform circulating gas lifting intelligent optimization and edge control collaborative method. BACKGROUND

[0002] In the middle and late stages of gas well development, gas well liquid loading is a key problem restricting production efficiency. As a core means to solve liquid loading, gas lifting and drainage gas recovery technology has formed a certain system, but there are still significant limitations in multi-well collaborative optimization, real-time dynamic regulation, etc.

[0003] The prior art has realized the breakthrough of single-well parameter design, process selection or single-point control in the field of gas lifting and drainage gas recovery, but it has not solved the problem of multi-well platform circulating gas lifting, and the multi-well coupling interference has not been effectively handled, lacking a platform-level gas injection collaborative optimization mechanism. The influence of formation gas production on gas injection has not been included in the calculation model, resulting in insufficient accuracy of gas injection. The "real-time data acquisition-edge side optimization-instruction execution" closed-loop control has not been formed, making it difficult to cope with dynamic changes in production parameters. Therefore, the present application proposes a platform circulating gas lifting intelligent optimization and edge control collaborative method to solve the problems in the prior art. SUMMARY

[0004] To solve the above problems, the present application proposes a platform circulating gas lifting intelligent optimization and edge control collaborative method. The platform circulating gas lifting intelligent optimization and edge control collaborative method draws a gas lifting dynamic curve considering formation gas production, establishes a multi-well collaborative optimization model, and constructs a closed-loop control system combined with edge computing technology to accurately determine the gas injection amount, eliminate multi-well coupling interference, realize intelligent optimization and real-time control of platform circulating gas lifting, and improve gas well production efficiency and gas field development benefit.

[0005] To achieve the purpose of the present application, the following technical solution is adopted: a platform circulating gas lifting intelligent optimization and edge control collaborative method, comprising the following steps:

[0006] Step one, single-well gas lifting dynamic curve drawing, calculating formation pressure and liquid production parameters, then fitting to obtain liquid production index and gas production index, and considering formation gas production, drawing a gas lifting dynamic curve with gas injection amount as the horizontal coordinate and liquid production and gas production as the vertical coordinate;

[0007] Step two, multi-objective optimization model establishment, taking maximum platform total gas production as the target, single-well gas injection upper and lower limits and platform total gas injection as the constraint conditions, constructing an optimization model expressed by the following formula:

[0008]

[0009] In the formula, q gi is the gas allocation of each well, i.e. gas injection amount, qi (q gi ) is the corresponding gas production of each well under the gas injection amount q gi , which does not include the gas injection amount, Q SumMax is the maximum value of the total gas production of the platform, and the units are both 10 4 m 3 / d, Q gTotal is the total gas injection amount of the platform, q gimin is the minimum gas injection amount of the i th well, q gimax is the maximum gas injection amount of the i th well, and the units are both 10 3 m 3 / d; p s is the separator pressure, P s is the set separator pressure, and the units are both MPa; Dest Fitness is the target fitness;

[0010] Step three, gas injection parameter optimization and dynamic adjustment, solving the model in step two, taking the gas injection scheme of multiple wells of the platform as one of the particles, using the particle swarm optimization algorithm to calculate the fitness of each particle for optimization and dynamic adjustment;

[0011] Step four, circulating gas lift intelligent control, building a circulating gas lift intelligent control system, collecting process and production data, then based on the particle swarm optimization algorithm, taking the maximum gas production as the target, real-time solving the multi-objective optimization model to obtain the optimal gas injection amount, then the optimized parameters are sent to the intelligent controller through the production network to adjust the gas lift valve opening, and the execution result is fed back in real time to complete the closed-loop control.

[0012] Further improvement lies in that: the formation pressure calculation in step one is specifically calculating the bottom hole flowing pressure through the static gas column gradient method according to the real-time casing pressure data, and then estimating the formation pressure, the calculation formula is formation pressure = bottom hole flowing pressure + 10 MPa, and the 10 MPa in the formula is the production pressure difference; the liquid production calculation is specifically calculated through the multiphase pipe flow calculation formula according to the real-time wellhead oil pressure, gas production and obtained bottom hole flowing pressure; the liquid production index and gas production index calculation is specifically to fit the liquid production index and gas production index according to two points (X, 0) and (Y, Z or M), wherein X is the formation pressure, Y is the bottom hole flowing pressure, Z is the liquid production, and M is the gas production.

[0013] Further improvement lies in that: the drawing of the gas lift dynamic curve in step one is specifically:

[0014] S1. Based on the gas well production index and gas production index, firstly calculate the maximum production volume using the liquid IPR method, and divide it into n equal parts to form a set of decreasing arithmetic series of production volumes Q1, ..., Qi, ..., Qn, where Q1 = Qmax and Qn = Qmax / n; based on the series of production volumes, calculate a series of corresponding bottom hole flowing pressures, and then calculate a series of corresponding gas production volumes Qg1, ..., Qgi, ..., Qgn using the bottom hole flowing pressure and gas IPR method;

[0015] S2. Based on a series of liquid production rates and corresponding gas production rates and corresponding bottom hole pressures, calculate from the bottom of the well to the injection point using multiphase flow to obtain a series of pressure values ​​at the injection point.

[0016] S3. Based on the given gas injection volume Qg, and considering the gas and liquid produced in S2, calculate the flowing pressure value from the gas injection point to the wellhead, i.e. the wellhead oil pressure, and obtain a series of wellhead oil pressures. According to the node analysis principle, calculate the liquid production and gas production corresponding to the wellhead oil pressure Pwh, and then obtain the liquid production and gas production corresponding to the gas injection volume Qg.

[0017] S4. Similarly, a series of liquid production and gas production rates corresponding to different gas injection rates can be obtained. With the gas injection rate as the horizontal axis and the liquid production rate and gas production rate as the vertical axis, the gas lift dynamic curve can be obtained.

[0018] The further improvement lies in the following: the fitness calculation in step three is specifically as follows:

[0019] S1. First, draw the inflow curve of each well with the wellhead as the node under different gas injection rates;

[0020] S2. For any solution The inflow curve at the wellhead of each gas lift well is obtained using interpolation.

[0021] S3. Take a node A on the inflow curve, when the pressure at node A is P. Ai The liquid flow rate q at node A of the gas lift well can be obtained from the inflow curve. ni (n = 1, 2, 3, ...), the liquid volume q at node A Ai =q 1i +q 2i +…+q ni From this, a series of (P) can be derived. A1 q A1 ), (P A2 q A2 )……(P Am q Am This yields the inflow curve for node A;

[0022] S4, the inflow curve of each node and the inflow curve of the separator as a node can be further obtained according to the set separator pressure P s The total liquid volume Q and the gas production can be obtained from the inflow curve of the separator, that is, the solution under the platform system;

[0023] S5, the solution under all platform systems is calculated, that is, the fitness of the particles.

[0024] Further improvement lies in that the optimization in step three comprises node analysis and well group optimization calculation, the node analysis specifically comprises setting the separator as a key node, establishing a multiphase pipe flow model from the wellhead to the separator, combining the IPR curve of each well with the pipe network equation, and solving the pressure and production coupling relationship; the well group optimization calculation specifically comprises automatically running the particle swarm optimization algorithm after importing the single well model, and outputting the optimized gas injection volume of each well;

[0025] The dynamic adjustment mechanism is that when the casing pressure changes by more than a set threshold, it is judged to be a liquid accumulation state or a part of liquid accumulation discharge state, that is, the well state changes dynamically, the formation pressure and the liquid / gas production index are recalculated, and the particle swarm optimization algorithm is triggered to update the gas distribution scheme.

[0026] Further improvement lies in that the circulating gas lifting intelligent control system in step four comprises a data acquisition and transmission module, an edge side optimization calculation module, an instruction execution and feedback module, and a cloud monitoring module; the data acquisition and transmission module collects process parameters and production data based on the interconnection of the RTU edge calculation terminal and the well station PLC system and the intelligent control valve; the edge side optimization calculation module constructs a multi-objective optimization model based on the PSO algorithm through the RTU edge calculation terminal, takes the maximum gas production as the target, and solves the optimal gas injection volume in real time; the instruction execution and feedback module is used to issue the optimized parameters to the intelligent controller through the production network, adjust the gas lifting valve opening degree, and feedback the execution result to the edge terminal in real time to form a closed loop control; and the cloud monitoring module is used to synchronize all data to the POC platform to realize remote monitoring and analysis.

[0027] Further improvement lies in that the specific control process in step four comprises:

[0028] S1, the RTU edge control terminal collects real-time data of oil pressure, casing pressure, gas nozzle pressure and temperature before and after throttling, instantaneous gas volume and instantaneous liquid volume;

[0029] S2, a large amount of real-time data is operated by the circulating gas lifting optimization algorithm module built in the RTU edge control terminal to solve the optimal gas injection volume;

[0030] S3, the injection amount is outside the prefabricated threshold, record the current data for experts and technical personnel to analyze, provide example data for subsequent algorithm optimization, the injection amount is within the threshold, send to the POC platform for confirmation or select automatic execution (test for a period of time and think it is reliable);

[0031] S4, in the non-automatic execution mode, after the POC confirms the optimized injection amount parameter, the RTU edge calculation terminal issues the optimized injection amount to the field intelligent control valve for injection optimization after the POC confirms that it can be issued;

[0032] S5, all data are synchronously uploaded to the POC platform to realize remote monitoring and analysis.

[0033] The beneficial effects of the present application are: the present application realizes the global optimal distribution of the injection amount through the two-stage optimization framework and the particle swarm algorithm, solves the coupling interference problem existing in the multi-well platform, realizes real-time response through edge control, avoids the production reduction and shutdown caused by liquid accumulation, realizes the precise control of the cyclic gas lifting process and the maximum yield target, and provides a reliable technical path for the intelligent drainage gas recovery of the shale gas field. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The present application is a method flowchart.

[0035] Figure 2 The present application is a gas lifting dynamic curve drawing flowchart.

[0036] Figure 3 The present application is a platform cyclic gas lifting particle swarm optimization flowchart.

[0037] Figure 4 The present application is a platform optimization scheduling scheme schematic diagram.

[0038] Figure 5 The present application is a gas lifting well production system architecture diagram.

[0039] Figure 6 The present application is an inflow curve diagram.

[0040] Figure 7 The present application is an optimized gas distribution result diagram.

[0041] Figure 8 The present application is a system data transmission flowchart.

[0042] Figure 9 The present application is a system control flowchart. DETAILED DESCRIPTION

[0043] In order to deepen the understanding of the present application, the present application will be further described below in conjunction with examples, which are only used to explain the present application and do not constitute a limitation on the protection scope of the present application.

[0044] According to Figures 1-9 The present embodiment provides a platform cyclic gas lifting intelligent optimization and edge control collaborative method, which comprises the following steps:

[0045] Step one, single well gas lifting dynamic curve drawing:

[0046] Calculate the formation pressure and liquid production parameters, then fit the liquid production index and gas production index, and consider the formation gas production, draw the gas lifting dynamic curve with the gas injection amount as the horizontal coordinate and the liquid production and gas production as the vertical coordinate;

[0047] The formation pressure calculation is specifically based on the real-time casing pressure data, the bottom hole flowing pressure is calculated through the static gas column gradient method, and then the formation pressure is estimated, and the calculation formula is formation pressure = bottom hole flowing pressure + 10 MPa, wherein 10 MPa is the production pressure difference;

[0048] The liquid production calculation is specifically based on the real-time wellhead oil pressure, gas production and obtained bottom hole flowing pressure, and is calculated through the multiphase pipe flow calculation formula;

[0049] The liquid production index and gas production index calculation is specifically based on the fitting of the two points (X, 0) and (Y, Z or M) to obtain the liquid production index and gas production index, wherein X is the formation pressure, Y is the bottom hole flowing pressure, Z is the liquid production, and M is the gas production;

[0050] The drawing process of the gas lifting dynamic curve is as shown in the accompanying drawings of the specification Figure 2 , and specifically is:

[0051] S1, according to the liquid production index and gas production index of the gas well, first calculate the maximum liquid production by using the liquid IPR method, and divide it into n equal parts to form a set of liquid production decreasing arithmetic sequences Q1, …Qi, …Qn, Q1=Qmax, Qn=Qmax / n; according to a series of liquid production, calculate a series of corresponding bottom hole flowing pressures, and then calculate a series of corresponding gas production Qg1, …Qgi, …Qgn according to the bottom hole flowing pressure and the gas IPR method;

[0052] S2, according to a series of liquid production and corresponding gas production, corresponding bottom hole flowing pressure, calculate the flow pressure values from the bottom to the gas injection point through multiphase flow, to obtain a series of flow pressure values at the gas injection point;

[0053] S3, according to the given gas injection amount Qg, considering the produced gas and liquid in S2, calculate the flow pressure value from the gas injection point to the wellhead, that is, the wellhead oil pressure, to obtain a series of wellhead oil pressures, according to the node analysis principle, the liquid production and gas production corresponding to the wellhead oil pressure Pwh are obtained, that is, the liquid production and gas production corresponding to the gas injection amount Qg are obtained.

[0054] S4. Similarly, a series of liquid production and gas production rates corresponding to different gas injection rates can be obtained. With the gas injection rate as the horizontal axis and the liquid production rate and gas production rate as the vertical axis, the gas lift dynamic curve can be obtained.

[0055] In actual production, the formation usually produces gas when gas lift is used to drain the liquid. Therefore, the gas injection rate needs to be optimized according to the actual formation gas production situation.

[0056] Step 2: Establishing a multi-objective optimization model:

[0057] Based on the analysis of the dynamic characteristic curve of gas lift, and considering the coupling interference effect of multiple wells on the platform, an intelligent gas distribution optimization model is constructed with the goal of maximizing gas production. Due to the influence of the dynamic balance of the formation-wellbore-pipeline system, each well has an optimal gas injection threshold, and exceeding this value will lead to a decrease in gas production. Based on this, a two-stage optimization framework is established: first, at the single-well level, the optimal gas injection threshold for each well is determined through IPR-VLP coupling analysis; second, at the platform level, each well is connected to the same separator through gathering and transportation pipelines. Changes in the gas injection or production of a certain well will cause fluctuations in the separator pressure Ps, which in turn will affect the wellhead oil pressure and bottomhole flowing pressure of other wells.

[0058] Using the nodal analysis principle, the separator is set as the key node, and a multiphase pipe flow model from the wellhead to the separator is established. By combining the inflow dynamic curve (IPR curve), outflow dynamic curve (VLP) of each well and the pipeline flow equation, the coupling relationship between pipeline pressure and the production of each well is solved.

[0059] When the total gas injection capacity of the platform is limited, increasing the gas injection capacity of a certain well may lead to insufficient gas injection pressure in other wells or exceed the total gas volume limit. Particle swarm optimization algorithm is introduced, with the gas injection capacity of each well as the particle dimension. With the goal of maximizing the total gas production of the platform, the optimal combination of gas injection capacity is searched through iteration, while satisfying the gas injection pressure limit of a single well. With the goal of maximizing the total gas production of the platform, the upper and lower limits of gas injection capacity of a single well and the total gas injection capacity of the platform are the constraints. The optimization model is constructed and expressed by the following formula.

[0060]

[0061] In the formula q gi The gas distribution rate for each well, i.e., the gas injection rate, q i (q gi ) represents the injection rate q for each well. gi The corresponding gas production volume below does not include the gas injection volume, Q. SumMax This represents the maximum total gas production of the platform, with units of 10. 4 m 3 / d, Q gTotal q represents the total gas injection volume of the platform. gimin Let q be the minimum gas injection rate for the i-th well.gimax Qi is the maximum gas injection rate of the ith well, unit is 10 3 m 3 / d;p s P is the separator pressure, P s Dest is the set separator pressure, unit is MPa; Dest Fitness F is the target fitness.

[0062] Step three, gas injection parameter optimization and dynamic adjustment:

[0063] Solve the model in step two, take the gas injection scheme of the platform wells as one of the particles, use particle swarm optimization algorithm to calculate the fitness of each particle for optimization and dynamic adjustment, the optimization process is shown in the attached Figure 3 ;

[0064] The most critical calculation is the calculation of the fitness of each particle; the calculation method of the particle fitness is as follows: for any optimization scheduling scheme of a platform in a time period, as shown in the attached Figure 4 , the production condition of each well corresponding to the scheme needs to be calculated according to the following calculation process.

[0065] The corresponding gas production and liquid production under the gas injection rate of each well, considering the mutual interference between gas lift wells, can be solved according to the principle of node analysis, using gas lift dynamic curve; for example: assuming that there is a gas lift well production system consisting of n wells and a separator, as shown in the attached Figure 5 .

[0066] Considering the mutual interference between wells and the ground gathering pipeline, the separator pressure is constant as a boundary condition. When the ground gathering pipeline pressure fluctuates, it will affect the gas well gas production and liquid production, and the influence relationship can be represented by the relationship curve between wellhead pressure and liquid production under different gas injection rates, as shown in the attached Figure 6 a, which is the wellhead inflow curve diagram of the ith well. According to the principle of node analysis, this curve is the inflow curve with the wellhead as the node.

[0067] The fitness calculation is as follows:

[0068] S1, first draw the inflow curve of each well with the wellhead as the node under different gas injection rates;

[0069] S2, for any solution Use interpolation method to solve the inflow curve of each gas lift well wellhead;

[0070] The fluid of node A comes from well 1 and well 2; for well 1, take a series of points on the inflow curve, calculate to node A according to the multiphase pipe flow theory, get the inflow curve of well 1 to node A, and similarly, get the inflow curve of well 2 to node A, as shown in the attachedFigure 6 Fig. 2b shows the inflow curve of well 1 and well 2 at node A;

[0071] S3, a node A is taken on the inflow curve, and when the pressure of node A is P Ai The liquid flow q ni of the gas-lift well inflow node A can be obtained from the inflow curve Ai (n = 1, 2, 3, …), the liquid flow q 1i of node A is q 2i + q ni , and thus a series of (P A1 , q A1 ), (P A2 , q A2 ) … (P Am , q Am ) can be obtained, i.e. the inflow curve of node A is obtained, as shown in Fig. 2c of the accompanying drawings; Figure 6 Fig. 2c shows the inflow curve of node A;

[0072] S4, the inflow curve of each node and the inflow curve of the separator as a node can be further obtained, as shown in Fig. 2d of the accompanying drawings; Figure 6 Fig. 2d shows the inflow curve of the separator; s The total liquid flow Q and the gas production can be obtained from the inflow curve of the separator, i.e. the solution under the platform system;

[0073] S5, the solutions under all platform systems are calculated, i.e. the fitness of the particles.

[0074] Finally, the optimized gas distribution result is formed, as shown in Fig. 3 of the accompanying drawings. Figure 7 Fig. 3 shows the optimized gas distribution result.

[0075] The optimization includes node analysis and well group optimization calculation:

[0076] Node analysis: the separator is set as the key node, the multiphase pipe flow model from the wellhead to the separator is established, the pressure and production coupling relationship is solved by combining the IPR curve of each well and the pipe network equation;

[0077] Well group optimization calculation: after the single well model is introduced, the particle swarm optimization algorithm is automatically run, and the optimized gas injection amount of each well is output.

[0078] The dynamic adjustment mechanism is that when the casing pressure changes by more than a set threshold, it is judged to be a liquid accumulation state, the formation pressure and the liquid / gas production index are recalculated, and the particle swarm optimization algorithm is triggered to update the gas distribution scheme.

[0079] Step four, intelligent control of gas lift:

[0080] The field has built a gas production process remote support center (POC) that can dynamically track the whole life cycle of gas well production, diagnose gas production process abnormalities and optimize parameters; in view of the intelligent control needs of the circulating gas lifting process of the shale gas platform, the system design needs to take into account the following core elements:

[0081] First, the centralized control level: through the intelligent controller to realize the unified regulation and control of multi-well gas lifting parameters; based on high-precision sensor network (pressure, flow, valve position, etc.) real-time monitoring of gas lifting efficiency, building a closed-loop feedback mechanism.

[0082] Second, the intelligent algorithm level: deploy an intelligent optimization model based on particle swarm optimization (PSO), and automatically generate optimization instructions for key parameters such as gas injection volume and gas injection timing in combination with real-time working conditions.

[0083] Build a circulating gas lifting intelligent control system, collect process and production data, then based on particle swarm optimization algorithm, maximize gas production as the goal, real-time solution of multi-objective optimization model, get the optimal gas injection volume, then the optimized parameters are sent to the intelligent controller through the production network, adjust the gas lifting valve opening, real-time feedback the execution result, complete the closed-loop control.

[0084] The circulating gas lifting intelligent control system includes a data acquisition and transmission module, an edge-side optimization calculation module, an instruction execution and feedback module, and a cloud monitoring module.

[0085] The data acquisition and transmission module collects process parameters (regulating valve opening, gas injection volume, etc.) and production data (instantaneous gas production, oil / suit pressure, etc.) through Modbus TCP / IP protocol and well station PLC system, intelligent control valve interconnection based on RTU edge computing terminal, along with the existing production network transmission channel, to ensure data compatibility and security;

[0086] The edge-side optimization calculation module constructs a multi-objective optimization model based on the PSO algorithm through the RTU edge computing terminal, maximizes gas production as the goal, and real-time solves the optimal gas injection volume;

[0087] The instruction execution and feedback module is used to send the optimized parameters to the intelligent controller through the production network, adjust the gas lifting valve opening, and real-time feedback the execution result to the edge terminal, forming a closed-loop control;

[0088] The cloud monitoring module is used to synchronize all data to the POC platform for remote monitoring and analysis.

[0089] The data transmission flow chart of the system is shown in the accompanying drawings Figure 8 The data to be collected by the whole intelligent control system and the implementation path are shown in Table 1.

[0090] Table 1 Data acquisition items and control devices

[0091]

[0092] The control flow of the system is shown in the accompanying drawings of the specification, and specifically includes: Figure 9

[0093] S1, the RTU edge control terminal carries out real-time data collection of oil, casing pressure, pressure and temperature before gas nozzle throttling, pressure and temperature after gas nozzle throttling, instantaneous gas volume and instantaneous liquid volume;

[0094] S2, a large amount of real-time data is calculated by the cyclic gas lifting optimization algorithm module built in the RTU edge control terminal, and the optimal gas injection volume is solved;

[0095] S3, if the gas injection volume is outside the pre-set threshold, the current data is recorded for analysis by experts and technical personnel, and example data is provided for subsequent algorithm optimization, and if the gas injection volume is within the threshold, it is sent to the POC platform for confirmation;

[0096] S4, after the POC confirms the optimized gas injection volume parameter, the RTU edge computing terminal sends it to the field intelligent control valve for gas injection optimization after confirmation;

[0097] S5, all data is synchronized and uploaded to the POC platform to realize remote monitoring and analysis.

[0098] The RTU edge computing terminal is composed of four parts: power module, communication module, control module and algorithm module;

[0099] The main equipment list is as follows:

[0100] Power supply PS1: 120W, input 220V, output 24V; including air circuit breaker QM1 and surge module SP1;

[0101] Communication module: 4-port industrial Ethernet switch SW1;

[0102] Control module and algorithm module: integrated RTU edge computing control device IOT1, including RTU module and 4Tops algorithm module.

[0103] The control principle is:

[0104] The connection of wellhead communication equipment and RTU on the physical layer is RS4J interface: the communication equipment data between multiple wells is received and sent through the application layer Modbus TCP protocol, achieving the purpose of bidirectional transmission of real-time data of multiple wells.

[0105] ​The RTU function items are researched, the RTU is based on an aarch64 kernel, a linux operating system platform, a cross-compilation tool aarch64-linux-gnu-g++ is used, an adaptation program is written to save real-time data received by RS485 in a database, and original data are provided for edge algorithm calculation; the libmodbus C library is called, a Modbus TCP host context program is written, register data of the RTU slave machine are read, then the sqlite3 C library is called, the original data are written into a.db database file, the edge algorithm program can read the database file, read the required data and perform corresponding calculation, the result after calculation is written into the database, the RTU receives the result, and issues an instruction to the wellhead control device to perform an action.

[0106] This time, the domestic RTU with edge computing capability is adopted, a linux-based algorithm environment is built, multi-well group fusion control optimization algorithm research is carried out, and a multi-working-condition operation mode of data local collection + algorithm local optimization + device local execution is realized.

[0107] Embodiment

[0108] In this embodiment, the AAA platform and the BBB platform are used to carry out cyclic gas lifting intelligent optimization control test, and the specific implementation is as follows:

[0109] I. AAA platform cyclic gas lifting intelligent optimization scheme and field implementation

[0110] (1) Gas well basic situation

[0111] There are 6 production wells in the AAA platform. In order to fully develop the gas production potential of each well in the platform, the "booster + gas lifting + foam drainage" fine drainage technology is planned to be put into operation. Before the operation of this technology, it is predicted that each well will increase gas production by 0-0.5 million square meters / MPa. The basic situation of each well is as follows:

[0112] ① Well AAA-1

[0113] Well AAA-1 was put into production in November 2021. In June 2022, 23 / 8" tubing was lowered under pressure (lower depth 2855m). In February 2023, the foam drainage technology was implemented. The initial daily gas production was 50,000 square meters, and the daily water production was 42 square meters. In the later period, the daily gas production gradually decreased to 30,000 square meters, the daily water production decreased to 3 square meters, and the gas lifting frequency was 1-5 times / year (since November 2024, there has been no gas lifting).

[0114] The well horizontal section is 2000 m long, EUR is 1.2 billion, cumulative production is 0.45 billion, current valve pressure is 2.2 MPa, casing pressure is 3.6 MPa, oil pressure is 2.4 MPa, daily gas production is 29,000 square meters, and daily water production is 3 square meters. Foam drainage production is used, and PIPESIM calculates the downhole liquid level of 0 m. According to the effect prediction of other platform pressure boosting wells implemented, the transmission pressure is reduced by 1.0 MPa, and the single well maximum production is increased by 0.5-0.6 thousand square meters. According to the effect prediction of the concentrated pressure boosting of the well in June 2023, the single well has no production effect.

[0115] ②Well AAA-2

[0116] Well AAA-2 was put into production in November 2021, and 23 / 8" tubing was lowered in June 2022 under pressure (lower depth 3129 m), and foam drainage technology was implemented in February 2023. The initial daily gas production is 57,000 square meters, and the daily water production is 30 square meters. Later, it gradually decreased to daily gas production of 31,000 square meters and daily water production of 2 square meters, and the gas lift frequency was 25-51 times / year (since November 2024, gas lift 2 times).

[0117] The well horizontal section is 1800 m long, EUR is 1.5 billion, cumulative production is 0.58 billion, current valve pressure is 2.2 MPa, casing pressure is 4.2 MPa, oil pressure is 2.4 MPa, daily gas production is 30,000 square meters, and daily water production is 2 square meters. The well uses foam drainage production, and PIPESIM calculates the downhole liquid level of 50 m. According to the effect prediction of the pressure boosting wells implemented, the transmission pressure is reduced by 1.0 MPa, and the single well production is increased by 0.5-0.6 thousand square meters. According to the effect prediction of the concentrated pressure boosting of the well in June 2023, the single well conservatively increases production by 0.05 thousand square meters / MPa.

[0118] ③Well AAA-4

[0119] Well AAA-4 was put into production in November 2021, and 23 / 8" tubing was lowered in June 2022 under pressure (lower depth 2878 m), and foam drainage technology was implemented in February 2023. The initial daily gas production is 74,000 square meters, and the daily water production is 28 square meters. Later, it gradually decreased to daily gas production of 33,000 square meters and daily water production of 2 square meters, and the gas lift frequency was 5-20 times / year (since November 2024, gas lift 10 times).

[0120] The well horizontal section is 1810 m long, EUR is 1.24 billion, cumulative production is 0.52 billion, current valve pressure is 2.2 MPa, casing pressure is 3.8 MPa, oil pressure is 2.3 MPa, daily gas production is 20,000 square meters, and daily water production is 2 square meters. The well uses foam drainage production, and PIPESIM calculates the downhole liquid level of 30 m. According to the effect prediction of the pressure boosting wells implemented, the transmission pressure is reduced by 1.0 MPa, and the single well production is increased by 0.5-0.6 thousand square meters. According to the effect prediction of the concentrated pressure boosting of the well in June 2023, the single well conservatively increases production by 0.15 thousand square meters / MPa.

[0121] ④Well AAA-5

[0122] Well AAA-5 was put into production in June 2022, and in August 2022, 23 / 8" tubing was lowered under pressure (depth 2922m), in April 2023, foam flooding technology was implemented, initial daily gas production was 62,000 cubic meters, and daily water production was 11 cubic meters, later gradually decreased to daily gas production of 34,000 cubic meters, daily water production of 3 cubic meters, and gas lift frequency of 5-11 times / year (since November 2024, gas lift 2 times).

[0123] The horizontal section of the well is 1600m long, EUR is 0.89 billion, cumulative production is 0.41 billion, current pressure behind the valve is 2.2MPa, casing pressure is 4.3MPa, oil pressure is 2.4MPa, daily gas production is 30,000 cubic meters, and daily water production is 2 cubic meters. The well is produced by foam flooding, PIPESIM calculates that the downhole liquid level is 60m, according to the effect prediction of the implemented booster well, the pressure drop is reduced by 1.0MPa, and the single well production is increased by 0.5-0.6 million cubic meters, according to the effect prediction of the single well concentrated booster in June 2023, the single well production is conservatively increased by 0.06 million cubic meters / MPa.

[0124] The platform officially put into operation the integrated fine production and drainage technology of booster gas lift and foam flooding on March 9, 2025, one month before the operation, the total gas production of the platform had decreased to 128,000 cubic meters / day, one month after the operation, the effect of increasing production was obvious, and the total gas production of the platform increased to 153,000 cubic meters / day. The execution of the technology is as follows:(1) booster: 4 wells are connected, the wellhead oil pressure is reduced from 2.0MPa to 1.1-1.3MPa;(2) gas lift: wells AAA-1, 4, and 5 are assisted by injecting 10-12 thousand cubic meters of gas per day; well AAA-2 is assisted by injecting 25,000 cubic meters of gas per day after being re-produced due to water flooding.(3) foam flooding: well AAA-2 uses composite foam flooding technology, each time 50kg is injected, and the dilution ratio is 1:6.

[0125] (2) Optimization scheme of circulating gas lift

[0126] According to the current production of wells AAA-1, 2, 4, and 5, the current formation pressures of the wells are estimated to be 19.4, 24.54, 19.87, and 19.36MPa respectively, the gas production indexes of the wells are 0.01, 0.0014, 0.0077, and 0.007 respectively, and the liquid production indexes of the wells are 0.3, 0.2, 0.1, and 0.2 respectively. The specific parameters of the wells are shown in the following table 2.

[0127] Table 2 Basic parameters of each well of the platform

[0128]

[0129] By drawing the single-well gas lift dynamic curve, the optimal gas injection rates of wells AAA-1, 2, 4, and 5 are calculated to be 0, 35, 15, and 20,000 cubic meters per day respectively.

[0130] (3) Application effect

[0131] According to the optimization simulation results, combined with the production situation of each well in the early stage and the current process operating conditions, the gas injection rate of AAA-2 and AAA-4 wells was increased to about 35,000 and 15,000 cubic meters per day respectively during April 17-22.

[0132] Through gas injection rate optimization, AAA-2 well resumed normal production, with average daily gas production increasing from 5,000 to 10,000 cubic meters; the average daily gas production of AAA-4 well remained at about 26,000 cubic meters, but the daily water production increased significantly, with a maximum increase of 4 cubic meters. During the gas injection rate optimization period, the total gas production of AAA platform increased from 137,000 to 148,000 cubic meters per day, an increase of 8%, and the gas lift parameter optimization took effect.

[0133] II. Intelligent optimization scheme and field implementation of BBB platform circulating gas lift

[0134] (1) Basic situation of gas wells

[0135] BBB platform has 4 production wells, which is the first deep shale gas platform to test the circulating gas lift process. Before the platform was put into operation with the circulating gas lift process, Wells 2 and 3 relied on the vehicle-mounted gas lift to maintain normal production; since the circulating gas lift process was put into operation on December 20, 2023, no vehicle-mounted gas lift has been used. At the initial stage of the process implementation, the average daily gas production of single well stabilized at 30-35 thousand cubic meters, but as the pressure and production further decreased, the total production of the platform had decreased to about 45 thousand cubic meters per day by May 2025.

[0136] ① Well BBB-1

[0137] Well BBB-1 was put into production in April 2022, and 27 / 8" tubing was lowered with pressure in February 2023 (lowered depth 75.7m), and the circulating gas lift process was implemented in January 2024. Before the process was put into operation, the casing pressure was 2.08-2.20 MPa, the oil pressure was 1.15-1.23 MPa, the daily gas production was 2.51-2.92 thousand cubic meters, the daily water production was 3-5 cubic meters, and almost no vehicle-mounted gas lift was used; after the process was put into operation, stable production was maintained, and by May 2025, the daily gas production had decreased to 14 thousand cubic meters.

[0138] ② Well BBB-2

[0139] Well BBB-2 was put into production in April 2022, and 27 / 8" tubing was lowered with pressure in February 2023 (lowered depth 70.6m), and the circulating gas lift process was implemented in December 2023. Before the process was put into operation, the casing pressure was 2.09-2.35 MPa, the oil pressure was 1.12-1.25 MPa, the daily gas production was 2.48-2.90 thousand cubic meters, the daily water production was 3-4 cubic meters, and the frequency of vehicle-mounted gas lift was 8.58 times per month; after the process was put into operation, stable production was maintained, and by May 2025, the daily gas production had decreased to 12 thousand cubic meters.

[0140] ③ Well BBB-3

[0141] Well BBB-3 was put into production in April 2022, and 27 / 8" tubing was lowered in March 2023 (lower depth 70.13m), and in December 2023, the circulating gas lifting process was implemented. Before the process was put into operation, the casing pressure was 2.33-2.76 MPa, the oil pressure was 1.19-1.27 MPa, the daily gas production was 1.11-4.90 million cubic meters, the daily water production was 0-5 cubic meters, and the frequency of vehicle-mounted gas lifting was 8.41 times / month. After the process was put into operation, stable production was maintained, and by May 2025, the daily gas production decreased to 60,000 cubic meters.

[0142] (4) Well BBB-4

[0143] Well BBB-4 was put into production in April 2022, and 27 / 8" tubing was lowered in March 2023 (lower depth 74.77m), and in January 2024, the circulating gas lifting process was implemented. Before the process was put into operation, the casing pressure was 2.76-2.82 MPa, the oil pressure was 1.18-1.27 MPa, the daily gas production was 3.81-4.77 million cubic meters, the daily water production was 3-6 cubic meters, and vehicle-mounted gas lifting was not used. After the process was put into operation, stable production was maintained, and by May 2025, the daily gas production decreased to 13,000 cubic meters.

[0144] (2) Optimization scheme of circulating gas lifting

[0145] Based on the existing facilities of BBB platform (separator, compressor, intelligent control valve, etc.), a new RTU edge computing terminal is added as the core control unit to build a closed-loop control system of "data acquisition-optimization calculation-instruction execution".

[0146] This test is divided into four steps:

[0147] Phase one (1-2 weeks): equipment installation and commissioning, install the RTU edge computing terminal in the BBB platform station skid house, and conduct network testing and joint commissioning with the gas lifting intelligent control equipment to ensure that the RTU edge computing terminal successfully reads the real-time production data and real-time gas injection parameters of the 4 wells on the BBB platform, and ensures that the gas injection amount instruction can be issued by the RTU edge computing terminal to control the gas injection amount of each well.

[0148] Phase two (3-4 weeks): manual optimization test, based on PSO algorithm to optimize single well gas injection amount, and verify the accuracy of the model by manually adjusting the gas injection amount.

[0149] Phase three (5-6 weeks): automatic optimization test, enable the circulating gas lifting intelligent control system to dynamically adjust the gas injection amount of the 4 wells, and compare the total production change before and after optimization.

[0150] Phase four: summary and evaluation, analyze and summarize the adaptability of the circulating gas lifting intelligent control technology.

[0151] (3) Application effect

[0152] Through integrated design, RTU edge computing terminal is integrated in a 4U standard longitudinal space, and is integrated and installed in the original cabinet. Through connecting standby UPS loop power supply and production network switch, system installation and connection are realized.

[0153] RTU edge computing terminal collects key real-time data such as station control PLC system oil jacket pressure, pressure before and after interception, temperature and gas injection amount of intelligent control valve of the platform, and the online real-time optimization result of cyclic gas lifting intelligent optimization algorithm can be issued to the intelligent control valve to realize centralized and unified control of the station process.

[0154] During manual control, according to the calculation result of the optimization model, the gas injection amount of BBB-1, 2, 3 and 4 wells is adjusted from 20,000, 18,000, 15,000 and 15,000 cubic meters per day to 4,000, 9,000, 16,000 and 10,000 cubic meters per day respectively. One day later, the total gas production of the platform increased from 45,000 cubic meters per day to 59,000 cubic meters per day, and the effect of increasing production was remarkable, but BBB-1 well appeared wellbore fluid loading. After increasing the gas injection amount of BBB-1 well to 20,000 cubic meters per day, the test gas injection amount was 10,000 cubic meters per day, and wellbore fluid loading appeared again. Due to the fact that the fluid loading was not eliminated, the gas injection amount was reduced, and the holiday system optimization was not timely, which caused serious fluid loading to water flooding and shutdown, and the gas injection amount of the well was increased to 35,000 cubic meters per day to resume production. Through manual control of gas injection, it is considered that the gas injection amount of BBB-1 well should not be less than 20,000 cubic meters per day, so as to ensure the normal production of gas well, a safety threshold control is set, that is, a lower limit of gas injection amount threshold is set, and when the gas injection amount calculated by the optimization result is lower than the lower limit of gas injection amount threshold, the lower limit of gas injection amount threshold is executed and an alarm is issued.

[0155] During automatic control, BBB-1, 2, 3 and 4 wells produced stably, the daily gas production of the platform decreased by 27%, the daily gas production of the platform was stable at more than 50,000 cubic meters per day, and the yield improvement target was achieved.

[0156] III. Summary

[0157] The reliability of the cyclic gas lifting optimization model and intelligent control system is verified through field test. Based on the actual parameters of BBB and AAA platforms, single well and multi-well collaborative cyclic gas lifting optimization model is constructed, and the optimal gas distribution scheme solved by particle swarm algorithm achieves remarkable results after field implementation: after optimizing the gas lifting gas injection amount of AAA0 and BBB platforms, the total gas production of the platforms is increased.

[0158] The optimized gas injection amount is highly consistent with the theoretical value of the gas lift dynamic curve (error <5%), which not only proves the reliability of the optimization model, but also verifies the feasibility of the cyclic gas lift process control system based on the intelligent algorithm in actual production. The test results fully demonstrate that through the establishment of an accurate multi-well coupled model and the use of an intelligent optimization algorithm, the precise control of the cyclic gas lift process and the maximization of the yield can be achieved, thereby providing a reliable technical path for the intelligent drainage gas recovery of shale gas fields.

[0159] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A collaborative method for intelligent optimization and edge control of platform circulating airlift, characterized in that, Includes the following steps: Step 1: Plot the dynamic curve of gas lift in a single well. Calculate the formation pressure and production parameters, then fit the production index and gas production index, and consider the formation gas production. Plot the dynamic curve of gas lift with the injection volume as the abscissa and the production volume and gas production volume as the ordinate. Step 2: Establish a multi-objective optimization model. With the goal of maximizing the total gas production of the platform, and the upper and lower limits of gas injection volume of a single well and the total gas injection volume of the platform as constraints, an optimization model is constructed, which is expressed by the following formula. In the formula q gi The gas distribution rate for each well, i.e., the gas injection rate, q i (q gi ) represents the injection rate q for each well. gi The corresponding gas production volume below does not include the gas injection volume, Q. SumMax This represents the maximum total gas production of the platform, with units of 10. 4 m 3 / d, Q gTotal q represents the total gas injection volume of the platform. gimin Let q be the minimum gas injection rate for the i-th well. gimax The maximum gas injection rate of the i-th well is expressed in units of 10. 3 m 3 / d;p s For separator pressure, P s The set separator pressure, all in MPa; Dest Fitness For target fitness; Step 3: Optimization and dynamic adjustment of gas injection parameters. Solve the model in Step 2. Take the gas distribution scheme of the platform's multi-well as one of the particles, and use the particle swarm optimization algorithm to calculate the fitness of each particle for optimization and dynamic adjustment. Step 4: Intelligent control of circulating gas lift. Build an intelligent control system for circulating gas lift, collect process and production data, and then, based on the particle swarm optimization algorithm, solve the multi-objective optimization model in real time with the goal of maximizing gas production to obtain the optimal gas injection rate. Subsequently, the optimized parameters are sent to the intelligent controller through the production network to adjust the opening of the gas lift valve and provide real-time feedback of the execution results to complete closed-loop control.

2. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: In step one, the formation pressure calculation is specifically based on real-time casing pressure data. The bottom hole flowing pressure is calculated using the static gas column gradient method, and then the formation pressure is estimated. The calculation formula is: Formation pressure = Bottom hole flowing pressure + 10MPa, where 10MPa is the production pressure difference. The production rate is calculated based on the real-time wellhead oil pressure, gas production rate, and bottom hole flowing pressure using a multiphase pipe flow calculation formula. The production index and gas production index are calculated by fitting two points (X, 0) and (Y, Z or M), where X is the formation pressure, Y is the bottom hole flowing pressure, Z is the production rate, and M is the gas production rate.

3. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: The specific steps for plotting the airlift dynamic curve in step one are as follows: S1. Based on the gas well production index and gas production index, firstly calculate the maximum production volume using the liquid IPR method, and divide it into n equal parts to form a set of decreasing arithmetic series of production volumes Q1, ..., Qi, ..., Qn, where Q1 = Qmax and Qn = Qmax / n; based on the series of production volumes, calculate a series of corresponding bottom hole flowing pressures, and then calculate a series of corresponding gas production volumes Qg1, ..., Qgi, ..., Qgn using the bottom hole flowing pressure and gas IPR method; S2. Based on a series of liquid production rates and corresponding gas production rates and corresponding bottom hole pressures, calculate from the bottom of the well to the injection point using multiphase flow to obtain a series of pressure values ​​at the injection point. S3. Based on the given gas injection volume Qg, and considering the gas and liquid produced in S2, calculate the flowing pressure value from the gas injection point to the wellhead, i.e. the wellhead oil pressure, and obtain a series of wellhead oil pressures. According to the node analysis principle, calculate the liquid production and gas production corresponding to the wellhead oil pressure Pwh, and then obtain the liquid production and gas production corresponding to the gas injection volume Qg. S4. Similarly, a series of liquid production and gas production rates corresponding to different gas injection rates can be obtained. With the gas injection rate as the horizontal axis and the liquid production rate and gas production rate as the vertical axis, the gas lift dynamic curve can be obtained.

4. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: The fitness calculation in step three is specifically as follows: S1. First, draw the inflow curve of each well with the wellhead as the node under different gas injection rates; S2. For any solution The inflow curve at the wellhead of each gas lift well is obtained using interpolation. S3. Take a node A on the inflow curve, when the pressure at node A is P. Ai The liquid flow rate q at node A of the gas lift well can be obtained from the inflow curve. ni (n = 1, 2, 3, ...), the liquid volume q at node A Ai =q 1i +q 2i +…+q ni From this, a series of (P) can be derived. A1 q A1 ), (P A2 q A2 )……(P Am q Am This yields the inflow curve for node A; S4. Similarly, the inflow curve for each node and the inflow curve with the separator as the node can be further obtained, based on the set separator pressure P. s The total liquid volume Q and gas production can be obtained from the inflow curve of the separator, which is the solution under this platform regime. S5. Calculate the solutions under all platform regimes, which are the fitness of the particles.

5. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: The optimization process in step three includes node analysis and well group optimization calculation. Node analysis specifically involves setting the separator as a key node, establishing a multiphase pipe flow model from the wellhead to the separator, combining the IPR curves of each well with the pipeline equations, and solving the coupling relationship between pressure and production. Well group optimization calculation specifically involves importing the single-well model and automatically running the particle swarm optimization algorithm to output the optimized gas injection rate for each well. The dynamic adjustment mechanism determines that when the casing pressure change exceeds a set threshold, it indicates a liquid accumulation state or a state after partial liquid discharge, meaning the well state has changed dynamically. In this case, the formation pressure and production / gas production index are recalculated, triggering the particle swarm optimization algorithm to update the gas distribution scheme.

6. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: The circulating gas lift intelligent control system in step four includes a data acquisition and transmission module, an edge-side optimization calculation module, an instruction execution and feedback module, and a cloud monitoring module. The data acquisition and transmission module collects process parameters and production data based on the interconnection between the RTU edge computing terminal and the well site PLC system and intelligent control valves. The edge-side optimization calculation module constructs a multi-objective optimization model based on the PSO algorithm through the RTU edge computing terminal, with the goal of maximizing gas production, and solves for the optimal gas injection rate in real time. The instruction execution and feedback module is used to send the optimization parameters to the intelligent controller through the production network, adjust the opening of the gas lift valve, and feed back the execution results to the edge terminal in real time to form a closed-loop control. The cloud monitoring module is used to synchronously upload all data to the POC platform to realize remote monitoring and analysis.

7. The platform circulating airlift intelligent optimization and edge control collaborative method according to claim 1, characterized in that: The specific control process in step four includes: S1, RTU edge control terminal performs real-time data acquisition of oil, casing pressure, pressure and temperature before throttling of air nozzle, pressure and temperature after throttling of air nozzle, instantaneous gas volume and instantaneous liquid volume. S2. A large amount of real-time data is processed by the circulating gas lift optimization algorithm module built into the RTU edge control terminal to solve for the optimal gas injection rate; S3. When the gas injection volume is outside the pre-set threshold, record the current data for analysis by experts and technicians, and provide example data for subsequent algorithm optimization. When the gas injection volume is within the threshold, send it to the POC platform for confirmation or select automatic execution. S4. In non-automatic execution mode, after the POC confirms the optimized gas injection parameters and confirms that they can be sent, the RTU edge computing terminal sends the data to the on-site intelligent control valve for gas injection optimization. S5. Synchronously upload all data to the POC platform to achieve remote monitoring and analysis.

Images