A throttling backpressure multistage control method based on multivariable intelligent decoupling

Abstract

The present invention discloses a multi-stage control method for throttling back pressure based on multi-variable intelligent decoupling, including selecting a single / multi-spool throttle valve mode of a throttle valve; calculating the pressure loss ΔP from the wellhead to the front end of the throttle valve; iL , Pressure loss ΔP from the rear end of the throttle valve to the outlet Lo , Required throttling pressure loss ΔP Ls , the throttling back pressure P required to be controlled Ls , the required real-time throttle valve flow area S Lc If it is a single spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area and the fuzzy position in the fuzzy control area; control the valve core S to the fuzzy position; use the neural network adaptive approximation method to adjust the valve core S from the fuzzy position to the precise position A1; if it is a multi-spool throttle valve mode, according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area; control the fuzzy valve core DA to the fuzzy position in the fuzzy control area; adopt the neural network adaptive approximation method to regulate the precise valve core DB from the fuzzy position to the precise position A1.

Description

Technical Field

[0001] The present invention relates to the technical field of managed pressure drilling for oil drilling, and in particular to a throttling back pressure multi-stage control method based on multivariable intelligent decoupling. Background Art

[0002] With increasing emphasis on safety in oil drilling operations, managed pressure drilling (MPD) and precision managed pressure drilling (MPD) technologies are being applied on a large scale in some oilfields due to their advantages in reducing drilling and well control risks and improving drilling efficiency. The principle of these technologies is to control bottomhole pressure through manual or automatic adjustment of wellhead casing pressure, achieving rapid equilibrium between wellbore pressure and formation pressure. This effectively addresses complex drilling issues such as lost circulation, kicks, and collapses caused by narrow safety density windows and multiple pressure systems.

[0003] Adjusting the wellhead casing pressure is a prerequisite for achieving the purpose of managed pressure drilling and fine managed pressure drilling. At present, the change of wellhead casing pressure is mainly achieved by manually or automatically adjusting the throttle valve to change the throttle back pressure. The existing throttle valve has four driving modes: manual, pneumatic, hydraulic and electric control. Managed pressure drilling and fine pressure control operations are mainly based on manual, hydraulic and electric control. The characteristics are as follows: hydraulic control drive is mainly used for straight-type throttle valves such as piston and needle types, which have good stability, low cost and poor timeliness; electric control drive is mainly used for rotary throttle valves such as orifice type, which have good stability, high precision, fast speed and high cost; manual control is suitable for straight-type and rotary throttle valves, which have high precision, good stability and low cost. The control effect is affected by changes in factors such as return fluid flow and solid phase content, and the timeliness is poor, and real-time automatic control cannot be achieved. Currently, both hydraulic and electric control modes can be automatically controlled through external structures. This allows for automatic control of the throttle valve to change the throttle back pressure, adjust the wellhead casing pressure, and ultimately control the bottomhole pressure to balance the formation pressure, achieving the goals of managed pressure drilling and precision managed pressure drilling. Therefore, how to quickly and accurately control the throttle back pressure is the key to achieving real-time and precise control of bottomhole pressure to balance formation pressure.

[0004] At present, the throttle valves that can realize automatic control in pressure-controlled drilling and fine pressure-controlled drilling are of single-spool type, and the control ability of the throttle valve is improved mainly through core algorithms (PID algorithm, prediction model algorithm, model-free adaptive algorithm) and flow structure (including valve core structure), but each has its limitations: PID algorithm, because of its simple algorithm, the effect and accuracy of its controller depend to a large extent on accurate parameter integration, and it is very difficult to adjust when encountering more complex situations such as nonlinearity and hysteresis, and it cannot support online modification; prediction model algorithm, due to geological uncertainty, formation heterogeneity and randomness of drilling construction, it is difficult to establish a more accurate prediction model, so the prediction model cannot meet the requirements of automatic control of pressure-controlled drilling; model-free adaptive algorithm, an adaptive algorithm that does not require the establishment of a model, can meet the flow control needs and is effective in standing pressure. , casing pressure, and flow rate all have nonlinearity, and the degree of nonlinearity will also change with the operating conditions. The use of a model-free adaptive algorithm can have the control capability to solve dynamic nonlinearity, and can realize manual and automatic commissioning, multi-target conversion, set value tracking and other actions, to ensure smooth and worry-free operation of the hydraulic throttle valve. It is more suitable for the requirements of the automatic control system of pressure-controlled drilling for the core algorithm, but it is poor in balancing the direct contradictory relationship between timeliness and accuracy, and the control timeliness and accuracy are affected by changes in factors such as the return fluid flow rate and solid content; the flow structure (including the valve core structure) is optimized, but it cannot adapt to large changes in fluid properties and flow parameters, and its adaptability range is poor. Under certain conditions, when the return flow fluctuates greatly, there may be problems such as difficulty in controlling the accuracy when the flow rate is small and high back pressure when the flow rate is large. Summary of the Invention

[0005] The present invention provides a throttling back-pressure multi-stage control method based on multi-variable intelligent decoupling to solve the problems existing in the prior art.

[0006] The technical solution adopted in the present invention is as follows:

[0007] The present invention discloses a throttling back pressure multi-stage control method based on multivariable intelligent decoupling, comprising:

[0008] Select the single / multi-spool throttle valve mode of the throttle valve, input the basic information of the throttle valve, and collect the throttle valve status information, ground pipeline information, fluid status and performance information, wherein the basic information of the throttle valve includes the flow coefficient C V The surface pipeline information includes the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve. iL , length L iL and the inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo , fluid properties and flow information include fluid density ρ, fluid viscosity μ, and fluid flow rate Q;

[0009] According to the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve iL , length L iL The inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo And the fluid density ρ, fluid viscosity μ, and fluid flow Q are used to calculate the pressure loss ΔP from the wellhead to the front end of the choke valve iL And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo ;

[0010] According to the pressure loss ΔP of the fluid from the wellhead to the front end of the throttle valve iL , Pressure loss ΔP from the rear end of the throttle valve to the outlet Lo and the required wellhead casing pressure P a Calculate the required throttling pressure loss ΔP Ls ;

[0011] According to the required throttling pressure loss ΔP Ls And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo Calculate the throttling back pressure P required for control Ls , and the required throttling pressure loss ΔP Ls , fluid density ρ, fluid flow Q and flow coefficient C q Calculate the required real-time throttle valve flow area S Lc ;

[0012] If it is a single spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area and the fuzzy position in the fuzzy control area;

[0013] Control the valve core S to the fuzzy position;

[0014] According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to control the valve core S from the fuzzy position to the precise position A1;

[0015] If it is a multi-spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area;

[0016] Control the fuzzy valve core DA to the fuzzy position of the fuzzy control area;

[0017] According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to adjust the precise valve core DB from the fuzzy position to the precise position A1.

[0018] In one possible implementation, ΔP iL , ΔP LoThe head loss model along the conventional pipe flow is used for calculation, and the relationship is:

[0019] ΔP iL =f(D iL , L iL ,ρ,μ,Q)

[0020] ΔP LO =f(D LO , L LO ,ρ,μ,Q)

[0021] Where ΔP iL is the pressure loss of the fluid from the wellhead to the front end of the choke valve, ΔP Lo D is the pressure loss of the fluid from the rear end of the throttle valve to the outlet, iL L is the inner diameter of the pipeline from the wellhead to the front end of the choke valve, iL D is the length of the pipeline from the wellhead to the front end of the choke valve, LO L is the inner diameter of the pipeline from the rear end of the throttle valve to the outlet, LO is the length of the pipeline from the rear end of the throttle valve to the outlet, ρ is the fluid density, μ is the fluid viscosity, and Q is the fluid flow rate;

[0022] Required throttling pressure loss ΔP Ls , the calculation formula is:

[0023] P a =ΔP iL +ΔP Ls +ΔP Lo

[0024] ΔP Ls =P a -(ΔP iL +ΔP Lo )

[0025] Where ΔP Ls Required throttling pressure loss, P a The required wellhead casing pressure.

[0026] In a feasible embodiment, the throttle back pressure P required to be controlled Ls and the required real-time throttle valve flow area S Lc The calculation formula is:

[0027] P Ls =ΔP Ls +ΔP Lo

[0028]

[0029]

[0030] Among them, d e is the equivalent diameter under the required throttling flow area, S Lc is the required real-time throttle valve flow area, ρ is the fluid density, C V is the flow coefficient, and Q is the fluid flow rate.

[0031] In one feasible implementation, a neural network adaptive approximation method is used to control the valve core S from the fuzzy position A1 to the precise position A1, including the anti-overrun lower limit correction:

[0032] Read the fuzzy position of the valve core S in real time;

[0033] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core S;

[0034] If the ratio ε<δ1, the valve core S is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction;

[0035] If the ratio ε≥δ1, the valve core S does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.5.

[0036] In a feasible implementation, a neural network adaptive approximation method is adopted to control the valve core S from the fuzzy position A1 to the precise position A1, and also includes an anti-overlimit correction:

[0037] Read the fuzzy position of the valve core S in real time;

[0038] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core S;

[0039] If the ratio ε≤δ2, the valve core S does not need to be adjusted;

[0040] If the ratio ε>δ2, the valve core S is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

[0041] In one feasible implementation, a neural network adaptive approximation method is used to control the precise valve core DB from the fuzzy position to the precise position A1, including the correction to prevent the lower limit from being exceeded:

[0042] Read the fuzzy position of the valve core DB in real time;

[0043] Calculate the required real-time throttle valve flow area S in real time LcThe ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core DB;

[0044] If the ratio ε<δ1, the valve core DA is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction;

[0045] If the ratio ε≥δ1, the valve core DA does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.5.

[0046] In a feasible implementation, a neural network adaptive approximation method is adopted to control the precise valve core DB from the fuzzy position to the precise position A1, and also includes an anti-lower limit correction:

[0047] Read the fuzzy position of the valve core DB in real time;

[0048] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core DB;

[0049] If the ratio ε≤δ2, the valve core DA does not need to be adjusted;

[0050] If the ratio ε>δ2, the valve core DA is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

[0051] In a feasible embodiment, according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area, including:

[0052] According to the required real-time throttle valve flow area S Lc Choose whether to select the fuzzy control area manually or automatically;

[0053] If manual is selected, set the number of fuzzy regions n, according to the required real-time throttle valve flow area S Lc Divide into n fuzzy areas;

[0054] The required throttle valve flow area S is calculated based on the real-time data of fluid density ρ and fluid flow Q. Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area;

[0055] If automatic is selected, the number of fuzzy regions n and the maximum flow area S of any fuzzy region are calculated. nmax ;

[0056] Set the number of fuzzy regions to n, where n = ceiling (S Lrrmax / S Lrmin );

[0057] Among them, S Lrmin 、S Lrmax The minimum and maximum throttle flow calculation areas required for the operating well are obtained by the maximum / minimum values of the fluid density ρ and the fluid flow Q. The required throttle flow area S is determined. Lr and S Lrmin 、S Lrmax It is obtained by the following formula;

[0058]

[0059]

[0060] According to the required throttle valve flow area S Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area.

[0061] In one feasible implementation, a neural network adaptive approximation method is adopted to control the valve core S from the fuzzy position A1 to the precise position A1, including:

[0062] According to the throttling back pressure P required for control Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure;

[0063] If |P Lm -P Ls |≤σ, the valve core S stops adjusting;

[0064] If |P Lm -P Ls |>σ, then calculate P respectively Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ;

[0065] Judgment P Lm -P Ls Value, if P Lm -P Ls <0, the valve core S is adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2;

[0066] If P Lm -P Ls >0, the valve core S is adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

[0067] In one feasible implementation, a neural network adaptive approximation method is adopted to control the precise valve core DB from the fuzzy position to the precise position A1, including:

[0068] According to the throttling back pressure P required for control Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure;

[0069] If |P Lm -P Ls |≤σ, the valve core DB stops adjusting;

[0070] If |P Lm -P Ls |>σ, then calculate P respectively Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ;

[0071] Judgment P Lm -P Ls Value, if P Lm -P L <0, then judge |S Lm -S Ls | / 2 value, if |S Lm -S Ls | / 2>S Lm -S n-1max , then recalculate the valve core DA to the new fuzzy area and the valve core DB to the new fuzzy position, the S n-1max The maximum flow area of the next level model area;

[0072] If |S Lm -S Ls | / 2≤S Lm -S n-1max , then the valve cores DA and DB are adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2;

[0073] If P Lm -P L >0, then judge |S Ls -S Lm | / 2 value, if |S Ls -S Lm | / 2>S nmax -S Ls , then recalculate the adjustment valve core DA to the new fuzzy area, and the valve core DB to the new fuzzy position, S nmax The maximum flow area of any fuzzy region;

[0074] If |S Ls -S Lm | / 2≤S nmax -S Ls , then the valve cores DA and DB are adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

[0075] The beneficial effects of the technical solution of the present invention are as follows:

[0076] The present invention solves the problem of difficulty in timely and accurate control of throttling back pressure caused by changes in fluid flow, fluid density, fluid viscosity, throttle valve flow structure, throttle valve flow area, flow coefficient, and set wellhead casing pressure, especially the difficulty in control caused by large fluctuations in drilling displacement and pressure build-up at the maximum flow area, through multi-variable coupling influence analysis of throttling back pressure, two-step multi-level control mode (fuzzy decoupling control, neural network decoupling control), dual-thread dual-objective operation, and fuzzy and adaptive precise composite control mode. It realizes multi-level adaptive control of throttling valve and improves the ability to quickly and accurately control throttling back pressure in pressure-controlled drilling and fine pressure-controlled drilling. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0078] Figure 1 Schematic diagram of multi-stage control flow based on single valve core and multi-valve core throttle valve (P a The wellhead casing pressure to be controlled can be set manually by manual calculation or automatically by computer software; S Lc To calculate the required flow area, it is reversely calculated through the throttling back pressure multivariable coupling influence model; S is the valve core of the single-spool throttle valve, which independently realizes fuzzy control and precise control; DA and DB are the fuzzy control valve core and precise control valve core of the dual-spool throttle valve, respectively;

[0079] Figure 2 Schematic diagram of fuzzy calculation process for setting the flow area of choke valve under wellhead casing pressure (Q is fluid flow rate, which is automatically collected by flow meter; ρ is fluid density; μ is fluid viscosity; P a is the wellhead casing pressure, ΔP Ls is the required throttling pressure loss, ΔP iL is the pressure loss of the fluid from the wellhead to the front end of the throttle valve; ΔP Lo C is the pressure loss of the fluid from the rear end of the throttle valve to the outlet; v is the flow coefficient; d e is the equivalent diameter of the throttle valve flow channel; S Lc To calculate the required real-time throttle valve flow area);

[0080] Figure 3 Schematic diagram of the correction process for the valve core S of a single-spool throttle valve to prevent the valve core from exceeding the lower limit of the fuzzy area (ε is the real-time flow area / the maximum flow area of the fuzzy control area; δ1 is the lower limit determination coefficient; S nmax is the maximum flowable area in any fuzzy region);

[0081] Figure 4 Schematic diagram of the spool S correction process for preventing the spool from exceeding the lower limit of the fuzzy region of a single spool throttle valve - the correction process for preventing the upper limit from exceeding the upper limit (δ2 is the upper limit determination coefficient);

[0082] Figure 5 Schematic diagram of the spool DA correction process for preventing the spool DB from exceeding the lower limit of the fuzzy area of the dual-spool throttle valve - correction process for preventing exceeding the lower limit;

[0083] Figure 6 Schematic diagram of the spool DA correction process for preventing the spool DB from exceeding the lower limit of the fuzzy area of the dual-spool throttle valve - the correction process for preventing the upper limit from exceeding;

[0084] Figure 7 Schematic diagram of the fuzzy area selection process (S Lr The real-time throttle valve flow area is calculated based on real-time data; n is the number of fuzzy regions; S Lrmin 、S Lrmax The minimum and maximum throttle valve flow calculation areas for the parameter ranges of drilling conditions and drilling fluid properties are set);

[0085] Figure 8 The neural network adaptive approximation control process of throttling back pressure in the fuzzy region of a single spool throttle valve (P L0 P is the throttling back pressure to be controlled calculated by setting the wellhead casing pressure; Lm is the throttle back pressure collected in real time; σ is the accuracy required for the settable throttle back pressure; S Lm 、S Ls PLm 、P L0 Required throttle valve flow area under throttling back pressure);

[0086] Figure 9 The neural network adaptive approximation control process of throttling back pressure in the fuzzy region of single spool throttle valve (S n-1max is the maximum flow area of the next level model area);

[0087] Figure 10a Schematic diagram of a single-spool throttle valve - divided into zones by rotation - the area is linear (1-8 are 8 fuzzy control zones);

[0088] Figure 10b Schematic diagram of a single-spool throttle valve - the valve core S is adjusted to position A1 in area 4 and then adaptively adjusted (A1 is the fuzzy control point);

[0089] Figure 10c Schematic diagram of a single-spool throttle valve - adjust the valve core S to C1 position. If it does not meet the requirements, continue to adjust (B1 is an auxiliary calculation point);

[0090] Figure 10d Schematic diagram of a single spool throttle valve - spool S adjusts D1 position and pauses adjustment when required (C1, D1 control points);

[0091] Figure 11a Schematic diagram of a dual-spool throttle valve - the area is evenly divided according to the spool rotation DA - the area is nonlinear (A2 is the fuzzy control point);

[0092] Figure 11b Schematic diagram of a dual-spool throttle valve - spool DA is adjusted to the fuzzy area 2 position;

[0093] Figure 11c Schematic diagram of a dual-spool throttle valve - spool DA is adjusted to position A2 and adaptive adjustment begins;

[0094] Figure 11d Schematic diagram of a dual spool throttle valve - adjust the C2 position of the spool DB, and if it does not meet the requirements, adjust it;

[0095] Figure 11e Schematic diagram of a dual-spool throttle valve - spool DB adjusts the D2 position and pauses adjustment when the requirements are met (B2 is the auxiliary calculation point, C2 and D2 are the control points, spool DA is the fuzzy control spool, and spool DB is the precise control spool);

[0096] Figure 12a Schematic diagram of the dual throttle valve spool DB ultra-fuzzy range adjustment - spool DA is adjusted to the fuzzy area 2 position;

[0097] Figure 12bThis is a diagram of the ultra-fuzzy range adjustment of the dual throttle valve spool DB - the spool S is adjusted to position A3 and begins adaptive adjustment (A3 is the fuzzy control point);

[0098] Figure 12c Schematic diagram for adjusting the ultra-fuzzy range of the dual throttle valve spool DB - adjust the spool DB to position C3, and adjust if it does not meet the requirements;

[0099] Figure 12d Schematic diagram for adjusting the ultra-fuzzy range of the dual throttle valve spool DB - if the fuzzy area exceeds the limit, the spools DA and D are adjusted (B3 is the auxiliary calculation point, C3 and D3 are the control points). DETAILED DESCRIPTION

[0100] The following embodiments are described in detail, with examples illustrated in the accompanying drawings. When the following description refers to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following embodiments are not intended to represent all possible implementations consistent with the present invention. They are merely examples of systems and methods consistent with certain aspects of the present invention as detailed in the claims.

[0101] In order to solve the problem that the timeliness and accuracy of throttling back pressure control are affected by the properties of the return fluid, flow parameters, throttle valve structure, etc. during the automatic pressure control operation of existing managed pressure drilling and fine managed pressure drilling, this paper proposes a multi-objective high-precision intelligent decoupling control algorithm based on intelligent decoupling control algorithms such as fuzzy decoupling control and neural network decoupling control by absorbing the advantages of PID algorithm, prediction model algorithm, and model-free adaptive algorithm, and establishing a multi-objective high-precision intelligent decoupling control algorithm suitable for throttling back pressure control of pressure control and fine managed pressure drilling. The core is a two-step, multi-level control algorithm, that is, the first step is based on fuzzy decoupling control. The first step is to achieve rapid coarse control through coupling control. The second step is to achieve adaptive fine control based on the neural network decoupling control approximation coupling term. Combining the characteristics and driving mechanism of single-spool and double-spool throttle valves, multi-target and multi-stage adaptive control of the throttle valve is realized, which reduces the disturbance influence of multivariable factors such as the properties of the return fluid and flow parameters, and improves the ability to quickly and accurately control the throttling back pressure, thereby improving the level of managed pressure drilling and fine managed pressure drilling operations, further reducing the risks of drilling operations and well control, and ensuring safe drilling in abnormal pressure formations, multiple complex formations, and formations with narrow safe density windows.

[0102] Therefore, the present invention proposes a multi-stage throttle backpressure control method based on multivariable intelligent decoupling. Based on intelligent decoupling control algorithms such as fuzzy decoupling control and neural network decoupling control, a multi-objective, high-precision intelligent decoupling control algorithm suitable for throttle backpressure control in pressure-controlled and fine-grained pressure-controlled drilling is established. Combining the characteristics and drive mechanisms of single-spool and dual-spool throttle valves, this algorithm implements multi-objective, multi-stage adaptive control of the throttle valves, reducing the perturbations of multivariable factors such as return fluid properties and flow parameters. Based on the throttle backpressure control requirements, rapid and high-precision control of the throttle backpressure is achieved, thereby improving the performance of pressure-controlled and fine-grained pressure-controlled drilling operations, further reducing drilling operation and well control risks, and ensuring safe drilling in abnormally pressured formations, complex formations, and formations with narrow safe density windows. The details are as follows.

[0103] See also Figures 1 to 9 .

[0104] The present invention provides a multi-level control method for throttling back pressure based on multi-variable intelligent decoupling, such as Figure 1 Shown, including:

[0105] S01: Select the single / multi-spool throttle valve mode of the throttle valve, input the basic information of the throttle valve, and collect the throttle valve status information, ground pipeline information, fluid status and performance information, wherein the basic information of the throttle valve includes the flow coefficient C V The surface pipeline information includes the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve. iL , length L iL and the inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo , fluid properties and flow information include fluid density ρ, fluid viscosity μ, and fluid flow rate Q.

[0106] S02: According to the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve iL , length L iL The inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo And the fluid density ρ, fluid viscosity μ, and fluid flow Q are used to calculate the pressure loss ΔP from the wellhead to the front end of the choke valve iL And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo , and the pressure loss ΔP of the fluid from the wellhead to the front end of the choke valve iL , Pressure loss ΔP from the rear end of the throttle valve to the outlet Lo and the required wellhead casing pressure P a Calculate the required throttling pressure loss ΔP Ls .

[0107] In one possible implementation, ΔP iL , ΔPLo The head loss model along the conventional pipe flow is used for calculation, and the relationship is:

[0108] ΔP iL =f(D iL , L iL ,ρ,μ,Q)

[0109] ΔP LO =f(D LO , L LO ,ρ,μ,Q)

[0110] Where ΔP iL is the pressure loss of the fluid from the wellhead to the front end of the choke valve, ΔP Lo D is the pressure loss of the fluid from the rear end of the throttle valve to the outlet, iL L is the inner diameter of the pipeline from the wellhead to the front end of the choke valve, iL D is the length of the pipeline from the wellhead to the front end of the choke valve, LO L is the inner diameter of the pipeline from the rear end of the throttle valve to the outlet, LO is the length of the pipeline from the rear end of the throttle valve to the outlet, ρ is the fluid density, μ is the fluid viscosity, and Q is the fluid flow rate;

[0111] Required throttling pressure loss ΔP Ls , the calculation formula is:

[0112] P a =ΔP iL +ΔP Ls +ΔP Lo

[0113] ΔP Ls =P a -(ΔP iL +ΔP Lo )

[0114] Where ΔP Ls Required throttling pressure loss, P a The required wellhead casing pressure.

[0115] S03: According to the required throttling pressure loss ΔP Ls And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo Calculate the throttling back pressure P required for control Ls , and the required throttling pressure loss ΔP Ls , fluid density ρ, fluid flow Q and flow coefficient C V Calculate the required real-time throttle valve flow area S Lc .

[0116] like Figure 2As shown, in one achievable embodiment, the throttling back pressure P required to be controlled Ls and the required real-time throttle valve flow area S Lc The calculation formula is:

[0117] P Ls =ΔP Ls +ΔP Lo

[0118]

[0119]

[0120] Among them, d e is the equivalent diameter under the required throttling flow area, S Lc is the required real-time throttle valve flow area, ρ is the fluid density, C V is the flow coefficient, and Q is the fluid flow rate.

[0121] S04: If it is a single spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area and the fuzzy position in the fuzzy control area.

[0122] Among them, it should be noted that

[0123] like Figure 7 As shown, in one feasible embodiment, according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area, including:

[0124] According to the required real-time throttle valve flow area S Lc Choose whether to select the fuzzy control area manually or automatically;

[0125] If manual is selected, set the number of fuzzy regions n, according to the required throttle valve flow area S Lc Divide into n fuzzy areas;

[0126] The required throttle valve flow area S is calculated based on the real-time data of fluid density ρ and fluid flow Q. Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area;

[0127] If automatic is selected, the number of fuzzy regions n and the maximum flow area S of any fuzzy region are calculated. nmax ;

[0128] Set the number of fuzzy regions to n, where n = ceiling (S Lrrmax / S Lrmin );

[0129] Among them, S Lrmin 、S Lrmax The minimum and maximum throttle flow calculation areas required for the operating well are obtained by the maximum / minimum values of the fluid density ρ and the fluid flow Q. The required throttle flow area S is determined. Lr and S Lrmin 、S Lrmax It is obtained by the following formula;

[0130]

[0131]

[0132] According to the required throttle valve flow area S Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area.

[0133] S05: Control valve core S to the fuzzy position.

[0134] S06: According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to control the valve core S from the fuzzy position to the precise position A1.

[0135] like Figure 3 As shown, in one feasible implementation, a neural network adaptive approximation method is adopted to control the valve core S from the fuzzy position A1 to the precise position A1, including the anti-lower limit correction:

[0136] Read the fuzzy position of the valve core S in real time;

[0137] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core S;

[0138] If the ratio ε<δ1, the valve core S is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction;

[0139] If the ratio ε≥δ1, the valve core S does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.5.

[0140] like Figure 4 As shown, in a feasible implementation, a neural network adaptive approximation method is adopted to control the valve core S from the fuzzy position A1 to the precise position A1, and also includes an anti-overlimit correction:

[0141] Read the fuzzy position of the valve core S in real time;

[0142] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core S;

[0143] If the ratio ε≤δ2, the valve core S does not need to be adjusted;

[0144] If the ratio ε>δ2, the valve core S is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

[0145] like Figure 8 and Figures 10a-10d As shown, in one feasible implementation, a neural network adaptive approximation method is adopted to control the valve core S from the fuzzy position A1 to the precise position A1, including:

[0146] According to the throttling back pressure P Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure;

[0147] If |P Lm -P Ls |≤σ, the valve core S stops adjusting;

[0148] If |P Lm -P Ls |>σ, then calculate P respectively Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ;

[0149] Judgment P Lm -P Ls Value, if P Lm -P Ls <0, the valve core S is adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2;

[0150] If P Lm -P Ls >0, the valve core S is adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

[0151] S07: If it is a multi-spool throttle valve mode, then according to the required real-time flow area SLc Determine the fuzzy control region.

[0152] S09: Control the fuzzy valve core DA to the fuzzy position in the fuzzy control area.

[0153] S09: According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to adjust the precise valve core DB from the fuzzy position to the precise position A1.

[0154] like Figure 5 As shown, in one feasible implementation, a neural network adaptive approximation method is adopted to control the precise valve core DB from the fuzzy position to the precise position A1, including the anti-overrun lower limit correction:

[0155] Read the fuzzy position of the valve core DB in real time;

[0156] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core DB;

[0157] If the ratio ε<δ1, the valve core DA is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction;

[0158] If the ratio ε≥δ1, the valve core DA does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.5.

[0159] like Figure 6 As shown, in a feasible implementation, a neural network adaptive approximation method is adopted to control the precise valve core DB from the fuzzy position to the precise position A1, and also includes an anti-lower limit correction:

[0160] Read the fuzzy position of the valve core DB in real time;

[0161] Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum flow area that can be adjusted to the fuzzy position of the valve core DB;

[0162] If the ratio ε≤δ2, the valve core DA does not need to be adjusted;

[0163] If the ratio ε>δ2, the valve core DA is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

[0164] like Figure 9 by, Figures 11a-11e as well as Figures 12a-12dAs shown, in one feasible implementation, a neural network adaptive approximation method is adopted to control the precise valve core DB from the fuzzy position to the precise position A1, including:

[0165] According to the throttling back pressure P Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure;

[0166] If |P Lm -P Ls |≤σ, the valve core DB stops adjusting;

[0167] If |P Lm -P Ls |>σ, then calculate P respectively Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ;

[0168] Judgment P Lm -P Ls Value, if P Lm -P L <0, then judge |S Lm -S Ls | / 2 value, if |S Lm -S Ls | / 2>S Lm -S n-1max , then recalculate the valve core DA to the new fuzzy area and the valve core DB to the new fuzzy position, the S n-1max is the maximum flow area of the next level fuzzy region;

[0169] If |S Lm -S Ls | / 2≤S Lm -S n-1max , then the valve cores DA and DB are adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2;

[0170] If P Lm -P L >0, then judge |S Ls -S Lm | / 2 value, if |S Ls -S Lm | / 2>S nmax -S Ls, then recalculate the adjustment valve core DA to the new fuzzy area, and the valve core DB to the new fuzzy position, S nmax The maximum flow area of any fuzzy region;

[0171] If |S Ls -S Lm | / 2≤S nmax -S Ls , then the valve cores DA and DB are adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

[0172] The present invention provides a throttling back pressure multi-stage control method based on multivariable intelligent coupling, which can realize adaptive multi-stage control of single-spool throttle valves and dual-spool throttle valves by controlling the throttle valve type through mode selection.

[0173] ① When controlling the existing single-spool throttle valve, a two-step, multi-level adaptive control system for the single-spool is adopted: the first step is to implement fuzzy control based on fuzzy control decoupling, that is, according to the wellhead casing pressure control requirements, the throttle valve flow area is inversely calculated, and then the spool fuzzy control position is quickly selected, and the spool S is controlled at the specified fuzzy control position (no repeated adjustment in the middle); the second step is to implement adaptive approximation precise control using neural network decoupling control, that is, the spool S is precisely controlled within the area where it is located using the adaptive approximation method. After the fuzzy control of the spool S is in place, the maximum flow area of the fuzzy control area is divided into each adjustment step in a nonlinear manner (divided into equal parts m, where m is any optional natural number that can be set by yourself, with a larger step when close to fully open and a smaller step when close to closed).

[0174] ② When controlling the dual-spool throttle valve, a two-step, multi-level adaptive control system is adopted for the dual-spool. The first step is to implement fuzzy control based on fuzzy control decoupling. According to the wellhead casing pressure control requirements, the throttle valve flow area is inversely calculated. Then, the spool fuzzy control area is quickly selected, and the spool DA is controlled to directly execute into the specified fuzzy control area (no repeated modulation in the middle). The second step is to implement adaptive approximation precise control using neural network decoupling control. The spool DB is precisely controlled within this area using adaptive approximation, and each adjustment step is divided in a nonlinear manner (the step size is large when the spool DB is close to full opening, and the step size is small when it is close to closing).

[0175] ③ When the real-time flow area / the maximum flow area of the fuzzy control area is less than δ1, δ1 is set to between 0 and 0.5: the adjusted fuzzy control area, position and distribution record fuzzy control level are calculated based on the current flow area, and the valve core S is driven to directly adjust to recalculate the position of the fuzzy control area. After it is in place, the neural network adaptive approximation control is started in the fuzzy area; the dual-spool throttle valve core DA automatically adjusts to the next level of fuzzy control level (the maximum flow area becomes smaller) and memorizes the flow area, and calculates the adjusted fuzzy control position based on the memorized flow area, and drives the valve core DB to directly adjust to this position. After it is in place, the neural network adaptive approximation control is started.

[0176] ④ When the real-time flow area / the maximum flow area of the fuzzy control area is greater than δ2, and δ2 is set between 0 and 0.5: the adjusted fuzzy control area, position and distribution record fuzzy control level are calculated based on the current flow area, and the valve core S is driven to directly adjust to recalculate the position of the fuzzy control area. After it is in place, the neural network adaptive approximation control is started within the fuzzy area; the dual-spool throttle valve core DA automatically adjusts to the next level of fuzzy control level (the maximum flow area becomes larger) and memorizes the flow area, and calculates the adjusted fuzzy control position based on the memorized flow area, and drives the valve core DB to directly adjust to this position. After it is in place, the neural network adaptive approximation control is started.

[0177] (2) The dual-spool throttle valve of the present invention is suitable for multi-stage control of pressure-controlled drilling: the valve core DA is used for fuzzy multi-stage control, and the valve core DB is used for regional adaptive precise control.

[0178] (3) The proposed multi-level control can realize n-level control (n is any selectable natural number) by manually or automatically determining and setting the control level of the valve core DA (valve core S for a single valve), i.e., the number of fuzzy control area setting areas, through the control system.

[0179] ① Manual setting is to set the fuzzy control area valve core according to the equal division method. The single valve core throttle valve divides the maximum flow area of each level area by the rotation degree into a linear relationship. The double valve core throttle valve divides the maximum flow area of each level area by the rotation degree into a nonlinear relationship.

[0180] ② Automatic setting reversely calculates the number of fuzzy control area settings and the maximum flow area of each area based on the throttling back pressure multivariable coupling influence model and the displacement range, maximum flow area of the throttle valve, and accuracy requirements set at the drilling site. The maximum flow area of each level area has a nonlinear relationship.

[0181] (4) A multivariable coupling influence model of throttling back pressure based on unsteady flow is proposed to provide a basis for the fuzzy control of valve core DA and valve core S: L=f(fluid flow rate Q, fluid density ρ, fluid viscosity μ, throttle valve flow area S L , flow coefficient C q ), where P L ,Q,ρ,μ,S L 、C q It is a real-time variable in drilling operation and throttling control. The required throttle valve flow area is reversely calculated through the coupling influence model, thereby obtaining the fuzzy control levels of valve core DA and valve core S.

[0182] (5) A dual-thread dual-objective composite control algorithm was established, which was used to achieve optimal matching control of the two valve cores of the dual-valve core throttle valve, thereby improving control timeliness and accuracy: the valve core DA fuzzy control target and the valve core DB adaptive control target were calculated using independent threads respectively. When the valve core DA control was in place, the two valve cores operated synchronously. When the real-time flow area / the maximum flow area of the fuzzy control area was less than δ1 or the real-time flow area / the maximum flow area of the fuzzy control area was greater than δ2, the valve core DA fuzzy control target and the valve core DB adaptive control were performed through independent thread calculation and analysis.

[0183] (6) A multi-level control method for throttling back pressure based on multi-variable intelligent decoupling is proposed. Through two-step control, fuzzy decoupling, neural network decoupling and other intelligent decoupling control algorithms are introduced to establish a throttling back pressure variable coupling influence model. Fuzzy control is achieved through the influence model, which solves the problem that the throttling back pressure is difficult to control in a timely and accurate manner due to changes in fluid flow, fluid density, fluid viscosity, fluid composition, throttling valve flow structure, throttling valve flow area, and set wellhead casing pressure, especially the problem of difficulty in controlling the throttling back pressure caused by large fluctuations in drilling displacement and pressure buildup at the maximum flow area. Adaptive and precise control is achieved by approximating the coupling term through a neural network.

[0184] The present invention solves the problem of difficulty in timely and accurate control of throttling back pressure caused by changes in fluid flow, fluid density, fluid viscosity, throttle valve flow structure, throttle valve flow area, flow coefficient, and set wellhead casing pressure, especially the difficulty in control caused by large fluctuations in drilling displacement and pressure build-up at the maximum flow area, through multi-variable coupling influence analysis of throttling back pressure, two-step multi-level control mode (fuzzy decoupling control, neural network decoupling control), dual-thread dual-objective operation, and fuzzy and adaptive precise composite control mode. It realizes multi-level adaptive control of throttling valve and improves the ability to quickly and accurately control throttling back pressure in pressure-controlled drilling and fine pressure-controlled drilling.

[0185] Example 1

[0186] like Figures 10a-10d A single spool throttle valve control is shown below:

[0187] Taking the semicircular orifice plate single valve core throttle valve as an example, assuming the maximum flow area is 80cm2 , the valve core S is divided into 8 fuzzy areas with a linear relationship of area: 10, 20, 30, 40, 50, 60, 70, 80 cm 2 .

[0188] Step 1: Select the single spool throttle valve mode and enter or select the throttle valve related structural parameters (standard throttle valves have fixed structural coefficients).

[0189] Step 2: Set the wellhead casing pressure P to be controlled a , and calculate the throttling back pressure P required for control Ls .

[0190] Step 3: Enter other relevant parameters.

[0191] Step 4: Calculate the required throttle valve flow area S Lc , the S calculated under this assumption is Lc =35cm 2 , then area 4 is the selected fuzzy area, and the fuzzy position is point A1.

[0192] Step 5: After the valve core S is controlled to point A1, the neural network adaptive approximation is adopted to start regulation, and the regulation range is area 4. Assume that P Ls =3.4MPa, but at this time the actual measured throttling back pressure P Lm =3.8MPa, and the back pressure control accuracy is σ=0.1MPa.

[0193] Step 6: |P Ls -P Lm |=0.4MPa>σ, then the valve core S needs to be adjusted.

[0194] Step 7: P Ls =3.4MPa, corresponding to an area of 35cm 2 ;P Lm =3.8MPa, the corresponding area is calculated to be 32cm 2 (Point B1) At this time, other parameters remain unchanged.

[0195] Step 8: Determine P Lm >P Ls , then the valve core S is adjusted in the direction of increasing the throttle valve flow area, and the step size is |S Ls -S Lm | / 2=1.5cm 2 , then the flow area at the position after adjustment is 33.5cm 2 (Point C1). Assume that the measured pressure here is 3.58 MPa.

[0196] Step 9: |P Ls -P Lm|=0.18MPa>σ, the valve core S needs to be adjusted.

[0197] Step 10: P Lm =3.58MPa, the corresponding area is calculated to be 33.5cm 2 . At this time, other parameters remain unchanged.

[0198] Step 11: Determine P Lm >P Ls , then the valve core S is adjusted in the direction of increasing the throttle valve flow area, and the step size is |S Ls -S Lm | / 2=0.75cm 2 , then the flow area at the position after adjustment is 34.25cm 2 (Point D1). Assume that the measured pressure here is 3.49 MPa.

[0199] Step 12: |P Ls -P Lm |=0.09MPa<σ, then the valve core S does not need to be adjusted. The above dependent variable parameters are assumed to have no fluctuations, no over-limit fluctuations, and the valve core S fluctuates within the range of area 4.

[0200] Step 13: Continue online real-time analysis. If |P appears again Ls -P Lm |>σ, there are changes in multivariate parameters, and repeat the second step.

[0201] Example 2

[0202] like Figures 11a-11e The dual spool control embodiment shown is as follows:

[0203] Taking the semicircular orifice plate type double valve core throttle valve as an example, assuming that the maximum flow area is 80cm 2 The valve core DA rotation degree is divided into 8 fuzzy areas, and the areas are nonlinearly related: 19.7, 38, 3, 53.5, 65.5, 73.5, 78, 79.7, 80 cm 2 .

[0204] Step 1: Select the dual-valve spool throttle valve mode and enter or select the throttle valve related structural parameters (for newly designed throttle valves, enter specific structural coefficients).

[0205] Step 2: Set the wellhead casing pressure P to be controlled a , and calculate the throttling back pressure P required for control Ls =3.4MPa.

[0206] Step 3: Enter other relevant parameters.

[0207] Step 4: Calculate the required throttle valve flow area S Lc , the S calculated under this assumption is Lc =35cm 2 , then area 2 is the selected fuzzy area, and the upper limit of the fuzzy area is point M.

[0208] Step 5: Control the valve core DA to point M and the valve core DB to point A2. After both valves are in place, they are synchronized and the neural network adaptive approximation is used to start regulating the valve core DB. The regulation range is area 2. Assume P Ls =3.4MPa, but at this time the actual measured throttling back pressure P Lm =3.8MPa, and the back pressure control accuracy is σ=0.1MPa.

[0209] Step 6: |P Ls -P Lm |=0.4MPa>σ, the valve core DB needs to be adjusted.

[0210] Step 7: P Ls =3.4MPa, corresponding to an area of 35cm 2 ;P Lm =3.8MPa, the corresponding area is calculated to be 32cm 2 (Point B2). At this time, other parameters remain unchanged.

[0211] Step 8: Determine P Lm >P Ls , then the valve core DB is adjusted in the direction of increasing the throttle valve flow area, with a step size of |S Ls -S Lm | / 2=1.5cm 2 , then the flow area at the position after adjustment is 33.5cm 2 (Point C2). Assume that the measured pressure here is 3.58 MPa.

[0212] Step 9: |P Ls -P Lm |=0.18MPa>σ, the valve core DB needs to be adjusted.

[0213] Step 10: P Lm =3.58MPa, the corresponding area is calculated to be 33.5cm 2 . At this time, other parameters remain unchanged.

[0214] Step 11: Determine P Lm >P Ls , then the valve core DB is adjusted in the direction of increasing the throttle valve flow area, with a step size of |S Ls -S Lm | / 2=0.75cm 2, then the flow area at the position after adjustment is 34.25cm 2 (Point D2). Assume that the measured pressure here is 3.49 MPa.

[0215] Step 12: |P Ls -P Lm |=0.09MPa<σ, then the valve core S does not need to be adjusted. The above dependent variable parameters are assumed to have no fluctuations, and no over-limit fluctuations occur, so the valve core DA does not need to be adjusted.

[0216] Step 13: Continue online real-time analysis. If |P appears again Ls -P Lm |>σ, there are changes in multivariate parameters, and repeat the second step.

[0217] Example 3

[0218] like Figures 12a-12d The hyper-fuzzy control range adjustment shown is as follows:

[0219] Taking the semicircular orifice plate type double valve core throttle valve as an example, assuming that the maximum flow area is 80cm 2 , the valve core S is divided into 8 fuzzy areas with a linear relationship of area: 10, 20, 30, 40, 50, 60, 70, 80 cm 2 .

[0220] Step 1: Select the dual-valve spool throttle valve mode and enter or select the throttle valve related structural parameters (for newly designed throttle valves, enter specific structural coefficients).

[0221] Step 2: Set the wellhead casing pressure P to be controlled a , and calculate the throttling back pressure P required for control Ls =3.4MPa.

[0222] Step 3: Enter other relevant parameters.

[0223] Step 4: Calculate the required throttle valve flow area S Lc , the S calculated under this assumption is Lc =35cm 2 , then area 3 is the selected fuzzy area, and the upper limit of the fuzzy area is point M.

[0224] Step 5: Control the valve core DA to point M, and control the valve core DB to point A3 at the same time. After both valves are in place, synchronize and use the neural network adaptive approximation to start regulating the valve core DB. The regulation range is area 3. Assume P Ls =3.4MPa, but at this time the actual measured throttling back pressure P Lm =3.8MPa, and the back pressure control accuracy is σ=0.1MPa.

[0225] Step 6: |P Ls -P Lm |=0.4MPa>σ, the valve core DB needs to be adjusted.

[0226] Step 7: P Ls =3.4MPa, corresponding to an area of 35cm 2 ;P Lm =3.8MPa, the corresponding area is calculated to be 32cm 2 (Point B3). Other parameters remain unchanged.

[0227] Step 8: Determine P Lm >P Ls , then the valve core DB is adjusted in the direction of increasing the throttle valve flow area, with a step size of |S Ls -S Lm | / 2=1.5cm 2 , then the flow area at the position after adjustment is 33.5cm 2 (Point C3). Assume that the measured pressure here is 3.58 MPa.

[0228] Step 9: |P Ls -P Lm |=0.18MPa>σ, the valve core DB needs to be adjusted.

[0229] Step 10: Multivariable parameters change, and the pressure control pressure P is set after calculation Ls Adjusted to 2.5MPa, the corresponding area is calculated to be 52cm 2 .

[0230] Step 11: Determine P Lm >P Ls , then the valve core DB is adjusted in the direction of increasing the throttle valve flow area, with a step size of |S Ls -S Lm | / 2=9.25cm 2 , then the flow area at the position after adjustment is 42.75cm 2 (point D3).

[0231] Step 12: After adjustment, the real-time flow area / the maximum flow area of the fuzzy control area = 42.75 / 38.0 = 1.13 > δ2, where the upper limit determination coefficient δ2 is set to 0.85.

[0232] Step 13: The fuzzy area needs to be adjusted by +1 level (increasing the area). After adjustment, the real-time flow area / the maximum flow area of the fuzzy control area = 42.75 / 53.5 = 0.80 < δ2. The fuzzy area is then determined to be 3 zones, and the fuzzy control position is calculated.

[0233] Step 14: Adjust valve core DA to fuzzy area 3, and simultaneously adjust valve core DB to fuzzy position D3. When both are in place, the two valve cores are synchronized, and valve core DB begins neural network adaptive approximation control to implement precise control. If over-limit occurs again, continue to adjust according to the above steps.

[0234] Similar parts between the embodiments provided in the present invention can be referenced to each other. The specific embodiments provided above are merely examples of the overall concept of the present invention and do not constitute a limitation on the scope of protection of the present invention. For those skilled in the art, any other embodiments expanded based on the scheme of the present invention without expending creative work shall fall within the scope of protection of the present invention.

Claims

1. A multi-level control method for throttling back pressure based on multi-variable intelligent decoupling, characterized in that: include: Select the single / multi-spool throttle valve mode of the throttle valve, input the basic information of the throttle valve, and collect the throttle valve status information, ground pipeline information, fluid status and performance information, wherein the basic information of the throttle valve includes the flow coefficient C V The surface pipeline information includes the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve. iL , length L iL and the inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo , fluid properties and flow information include fluid density ρ, fluid viscosity μ, and fluid flow rate Q; According to the inner diameter D of the pipeline from the wellhead to the front end of the throttle valve iL , length L iL The inner diameter D of the pipeline from the rear end of the throttle valve to the outlet Lo , length L Lo And the fluid density ρ, fluid viscosity μ, and fluid flow Q are used to calculate the pressure loss ΔP from the wellhead to the front end of the choke valve iL And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo ; According to the pressure loss ΔP of the fluid from the wellhead to the front end of the throttle valve iL , Pressure loss ΔP from the rear end of the throttle valve to the outlet Lo and the required wellhead casing pressure P a Calculate the required throttling pressure loss ΔP Ls ; According to the required throttling pressure loss ΔP Ls And the pressure loss ΔP from the rear end of the throttle valve to the outlet Lo Calculate the throttling back pressure P required for control Ls , and the required throttling pressure loss ΔP Ls , fluid density ρ, fluid flow Q and flow coefficient C V Calculate the required real-time throttle valve flow area S Lc ; If it is a single spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area and the fuzzy position in the fuzzy control area; Control the valve core S to the fuzzy position; According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to control the valve core S from the fuzzy position to the precise position A1; If it is a multi-spool throttle valve mode, then according to the required real-time throttle valve flow area S Lc Determine the fuzzy control area; Control the fuzzy valve core DA to the fuzzy position of the fuzzy control area; According to the throttling back pressure P required for control Ls The neural network adaptive approximation method is used to control the precise valve core DB from the fuzzy position to the precise position A1; according to the throttling back pressure P Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure; If |P Lm -P Ls |≤σ, the valve core S stops adjusting; If |P Lm -P Ls |>σ, then inversely calculate P Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ; Judgment P Lm -P Ls Value, if P Lm -P Ls <0, the valve core S is adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2; If P Lm -P Ls >0, the valve core S is adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

2. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 1 is characterized in that: ΔP iL , ΔP Lo The head loss model along the conventional pipe flow is used for calculation, and the relationship is: ΔP iL =f(D iL ,L iL ,p,m,Q) ΔP Lo =f(D Lo ,L Lo ,p,m,Q) Where ΔP iL is the pressure loss of the fluid from the wellhead to the front end of the choke valve, ΔP Lo D is the pressure loss of the fluid from the rear end of the throttle valve to the outlet, iL L is the inner diameter of the pipeline from the wellhead to the front end of the choke valve, iL D is the length of the pipeline from the wellhead to the front end of the choke valve, Lo L is the inner diameter of the pipeline from the rear end of the throttle valve to the outlet, Lo is the length of the pipeline from the rear end of the throttle valve to the outlet, ρ is the fluid density, μ is the fluid viscosity, and Q is the fluid flow rate; Required throttling pressure loss ΔP Ls , the calculation formula is: P a =ΔP iL +ΔP Ls +ΔP Lo ΔP Ls =P a -(ΔP iL +ΔP Lo ) Where ΔP Ls Required throttling pressure loss, P a The required wellhead casing pressure.

3. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 2 is characterized in that: The throttling back pressure P required for control Ls and the required real-time throttle valve flow area S Lc The calculation formula is: P Ls =ΔP Ls +ΔP Lo Among them, d e is the equivalent diameter under the required throttling flow area, S Lc is the required real-time throttle valve flow area, ρ is the fluid density, C V is the flow coefficient, and Q is the fluid flow rate.

4. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 1 or 3, characterized in that: The neural network adaptive approximation method is used to control the valve core S from the fuzzy position A1 to the precise position A1, including the correction to prevent the lower limit: Read the fuzzy position of the valve core S in real time; Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε to the maximum adjustable flow area in the fuzzy control area where the valve core S is located; If the ratio ε<δ1, the valve core S is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction; If the ratio ε≥δ1, the valve core S does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.

5.

5. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 4 is characterized in that: The neural network adaptive approximation method is used to control the valve core S from the fuzzy position A1 to the precise position A1, and also includes the correction to prevent the upper limit from exceeding: Read the fuzzy position of the valve core S in real time; Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε to the maximum adjustable flow area in the fuzzy control area where the valve core S is located; If the ratio ε≤δ2, the valve core S does not need to be adjusted; If the ratio ε>δ2, the valve core S is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

6. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 1 or 3, characterized in that: The neural network adaptive approximation method is used to control the precise valve core DB from the fuzzy position to the precise position A1, including the correction to prevent the lower limit: Read the fuzzy position of the valve core DB in real time; Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum adjustable flow area in the fuzzy control area where the valve core DB is located; If the ratio ε<δ1, the valve core DA is adjusted to level -1, and the fuzzy position is in the area S nmax area reduction; If the ratio ε≥δ1, the valve core DA does not need to be adjusted, and δ1 is the over-lower limit determination coefficient, and the range of δ1 is 0-0.

5.

7. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 6 is characterized in that: The neural network adaptive approximation method is used to control the precise valve core DB from the fuzzy position to the precise position A1, including the correction to prevent over-limit: Read the fuzzy position of the valve core DB in real time; Calculate the required real-time throttle valve flow area S in real time Lc The ratio ε of the maximum adjustable flow area in the fuzzy control area where the valve core DB is located; If the ratio ε≤δ2, the valve core DA does not need to be adjusted; If the ratio ε>δ2, the valve core DA is adjusted to level +1, and the fuzzy position is in the area S nmax The area increases, the δ2 is the upper limit determination coefficient, and the range of δ2 is 0.5-1.

8. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 1 is characterized in that: According to the required real-time throttle valve flow area S Lc Determine the fuzzy control area, including: According to the required real-time throttle valve flow area S Lc Choose whether to select the fuzzy control area manually or automatically; If manual is selected, set the number of fuzzy regions n, according to the required real-time throttle valve flow area S Lc Divide into n fuzzy areas; The required throttle valve flow area S is calculated based on the real-time data of fluid density ρ and fluid flow Q. Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area; If automatic is selected, the number of fuzzy regions n and the maximum flow area S of any fuzzy region are calculated. nmax ; Set the number of fuzzy regions to n, where n = ceiling (S Lrrmax / S Lrmin ); Among them, S Lrmin 、S Lrmax The minimum and maximum throttle flow calculation areas required for the operating well are obtained by the maximum / minimum values of the fluid density ρ and the fluid flow Q. The required throttle flow area S is determined. Lr and S Lrmin 、S Lrmax It is obtained by the following formula; According to the required throttle valve flow area S Lr With the maximum flow area S nmax The fuzzy area with the ratio closest to 50% is selected as the fuzzy control area.

9. The throttling back pressure multi-stage control method based on multivariable intelligent decoupling according to claim 1 is characterized in that: The neural network adaptive approximation method is used to control the precise valve core DB from the fuzzy position to the precise position A1, including: According to the throttling back pressure P required for control Ls Value, obtain the real-time collected throttling back pressure P Lm and the required throttling back pressure P calculated by setting the wellhead casing pressure Ls Value, Get|P Lm -P Ls |, and set the required control accuracy σ of throttling back pressure; If |P Lm -P Ls |≤σ, the valve core DB stops adjusting; If |P Lm -P Ls |>σ, then inversely calculate P Lm 、P Ls The corresponding required throttle valve flow area S Lm 、S Ls ; Judgment P Lm -P Ls Value, if P Lm -P L <0, then judge |S Lm -S Ls | / 2 value, if |S Lm -S Ls | / 2>S Lm -S n-1max , then recalculate the valve core DA to the new fuzzy area and the valve core DB to the new fuzzy position, the S n-1max The maximum flow area of the next level model area; If |S Lm -S Ls | / 2≤S Lm -S n-1max , then the valve cores DA and DB are adjusted in the direction of reducing the required throttle valve flow area, with a step size of |S Lm -S Ls | / 2; If P Lm -P L >0, then judge |S Ls -S Lm | / 2 value, if |S Ls -S Lm | / 2>S nmax -S Ls , then recalculate the adjustment valve core DA to the new fuzzy area, and the valve core DB to the new fuzzy position, S nmax The maximum flow area of any fuzzy region; If |S Ls -S Lm | / 2≤S nmax -S Ls , then the valve cores DA and DB are adjusted in the direction of increasing the required throttle valve flow area, with a step size of |S Ls -S Lm | / 2.

Images