A method and program product for dynamically controlling wellbore pressure in natural gas hydrate drilling

Abstract

The present invention discloses a method and program product for dynamic control of wellbore pressure in natural gas hydrate drilling, which establishes an original matrix of observation parameters and an original matrix of state parameters, performs data normalization processing of hysteresis regression and missing completion on the observation parameters based on adjacent well and drilling data; combines dual gradient drilling technology to establish a dynamic control model of wellbore pressure in drilling; adopts a particle filter algorithm to predict state parameters, and uses statistical errors to evaluate the prediction effect; establishes dual constraints of drilling safety density window and overflow safety evaluation standard to verify the safety of state parameters, outputs data if the conditions are met, and updates the original matrix of state parameters to a sequence matrix of state parameters. The present invention can predict the wellbore pressure profile at any time during the natural gas hydrate drilling process, improves the calculation accuracy of multiphase flow under complex constraints, and can increase the maximum drilling depth of a single opening by optimizing the observation parameters, thereby ensuring safe and efficient drilling.

Description

Technical Field

[0001] The invention belongs to the field of marine drilling engineering, and in particular relates to a natural gas hydrate drilling wellbore pressure dynamic control method and program product. Background Art

[0002] Natural gas hydrate is an energy resource with widespread reservoir distribution and huge potential reserves. It is considered to be an important alternative to clean energy in the future. However, due to its generally weakly consolidated and undiagenetic reservoirs, natural gas hydrate drilling faces many challenges such as wellbore instability, shallow high-pressure gas invasion and a narrow safe density window.

[0003] Dynamic wellbore pressure control technology is an effective means to solve the complex drilling problems of natural gas hydrates. It accurately predicts the wellbore pressure, quickly corrects and handles the changes in bottom hole pressure, controls the pressure fluctuations within a safe range, and avoids the influx of formation fluids. This is the key to meeting the needs of safe and efficient natural gas hydrate drilling.

[0004] Patent publication number CN116974312A discloses a method for controlling wellbore pressure in natural gas hydrate drilling and production, which realizes pressure prediction through pressure matching curve, but is not suitable for formations with narrow safety density window; Patent publication number CN113806945A discloses a formation inversion method for pressure-controlled drilling and production, which uses unscented Kalman filtering to invert gas production index and formation pressure, but this method has poor adaptability to non-Gaussian nonlinear problems.

[0005] Combining the actual needs of field applications and the shortcomings of the above patents, it is concluded that the current methods for controlling wellbore pressure in natural gas hydrate drilling have the following four problems:

[0006] (1) The raw data lacks standardized processing. The data required for wellbore pressure prediction comes from various sensors. Such data has lag. Taking bottom hole pressure as an example, the MWD is installed inside or above the drill collar, not at the bottom of the well. If the lag is 5 meters, when there is a drilling speed difference of 0.52 to 183.5 meters per hour, the bottom hole pressure obtained has a lag of about 0.03 to 9.62 hours. At the same time, high-frequency vibration and low temperature working conditions are also prone to data transmission failures. Data loss often occurs in actual drilling. If the lag time and data loss are not considered, it will cause large calculation errors;

[0007] (2) The prediction model is not suitable for drilling characteristics. Due to the secondary decomposition of gas hydrate cuttings, there is a sudden change in wellbore pressure. It is necessary to model the gas invasion. Different drilling methods directly affect the pressure control process. Conventional models do not combine the common drilling technology of gas hydrates. At the same time, the parameters of conventional models are only 1 to 2, which cannot accurately invert the overall trend of wellbore pressure.

[0008] (3) The inversion algorithm has a large error. The flow state of natural gas hydrate drilling wellbore is a multiphase flow with complex constraints. The annular fluid-solid coupling effect is obvious. The prediction method belongs to a typical non-Gaussian nonlinear problem. At present, the inversion algorithm is mainly based on extended Kalman filtering and unscented Kalman filtering. The theoretical accuracy is not high when dealing with non-Gaussian nonlinear problems.

[0009] (4) Safety verification is out of touch with reality. Dynamic control of wellbore pressure is meaningful only when it is achieved under the premise of safe drilling. However, some methods lack the necessary safety verification or constraint restrictions, and cannot ensure that the calculated parameters meet the safety density window and overflow safety requirements at the same time, and cannot safely guide production applications. Summary of the invention

[0010] In order to solve the above-mentioned drawbacks of the prior art, the present invention discloses a method and program product for dynamically controlling wellbore pressure in natural gas hydrate drilling.

[0011] The method for dynamically controlling the wellbore pressure of natural gas hydrate drilling comprises the following steps:

[0012] S100, establishing an original matrix of observation parameters and an original matrix of state parameters, and performing data normalization processing of lag regression and missing complement on the observation parameters based on adjacent wells and while-drilling data, to form a pre-processed observation parameter sequence matrix;

[0013] S200, establishing a dual-gradient pressure-controlled drilling prediction method and a gas invasion condition overflow estimation method to form a drilling wellbore pressure dynamic control model;

[0014] S300, establishing observation equations and state equations, using particle filter algorithm, combined with the drilling wellbore pressure dynamic control model, predicting state parameters based on observation parameters, and evaluating the prediction effect of state parameters using statistical errors;

[0015] S400, verify the safety of the state parameters according to the drilling safety density window restriction and the overflow safety evaluation standard restriction, output data if the conditions are met, and update the original matrix of the state parameters to the state parameter sequence matrix.

[0016] Furthermore, step S100 is specifically as follows:

[0017] S101, establish the original matrix of observation parameters, with a total of 10 observation parameters, divided into dual-gradient pressure-controlled drilling observation parameters and gas invasion overflow observation parameters:

[0018] {y}_{e}=[\, {p}_{d}\, \, {\rho}_{ECD}\, \, \rho \, \, {p}_{w}\, \, H\, \, {p}_{f}\, ] (6);

[0019] {y}_{h}=[\, T\, \, {p}_{a}\, \, \Delta {Q}_{p}\, \, {\rho}_{a}\, ] (7);

[0020] In formula (6), It is the observation parameter of dual gradient managed pressure drilling; is the logging while drilling pressure data, unit: MPa; is the equivalent circulating density of drilling fluid, in kg / m 3 ; is the mixed density of drilling fluid, in kg / m 3 ; is the wellhead back pressure, in MPa; is the well depth, in m; is the cyclic friction, in MPa;

[0021] In formula (7), It is the overflow observation parameter of gas intrusion condition; is the wellbore temperature, in °C; is the annulus pressure, in MPa; is the wellhead flow difference, in L / s; is the additional density of cuttings, in kg / m 3 ;

[0022] S102, establish the original matrix of state parameters, with a total of 4 observation parameters, divided into dual-gradient pressure-controlled drilling state parameters and gas invasion overflow state parameters:

[0023] {x}_{e}=[\, {p}_{b}\, \, {H}_{x}\, ] (8);

[0024] {x}_{h}=[\, {p}_{s}\, \, {k}_{r}\, ] (9);

[0025] In formula (8), It is the state parameter of dual gradient managed pressure drilling; is the predicted bottom hole pressure, in MPa; is the variable liquid level height of dual gradient drilling fluid, in m;

[0026] In formula (9), It is the overflow state parameter of gas intrusion condition; is the formation pressure, in MPa; is the gas intrusion rate, in m 3 / s;

[0027] S103, based on the adjacent wells and while-drilling data, the observed parameters are subjected to data normalization processing of lag regression and missing data completion. Specifically, on the basis of considering the late arrival time factors of each observed parameter, the neighbor interpolation method and the complex-valued function fitting method are used to perform lag regression of the observed parameters, and to predict and fill in the missing data, so as to eliminate the dimensional influence between different attributes;

[0028] (10);

[0029] (12);

[0030] (13);

[0031] In formula (10), is the normalized value of the i-th observed parameter; It is the minimum value of the normalized value of the observed parameter within the calculation step; It is the maximum value of the normalized value of the observed parameter within the calculation step; It is the minimum value of the original value of the observed parameter within the calculation step; It is the maximum value of the original value of the observed parameter within the calculation step;

[0032] In formula (12), is the characteristic coefficient, dimensionless; is the total number of samples of the observed parameter, in units; is the original value of the ith observed parameter. When the observed parameter is missing, the data of the adjacent well at the same location is taken; is the average value of the original value of the observed parameter; is the average of the normalized values ​​of the observed parameters; is the distribution factor, the initial value is 1;

[0033] In formula (13), is the hysteresis coefficient, dimensionless; is the borehole diameter, in mm; is the drill diameter, in mm; is the pump displacement, in L / s; is the mechanical drilling speed, in m / s; is the distance between the sensor and the calculation point, in meters; is the process factor, and its initial value is 1;

[0034] S104, introduced The fitting error is The fitting error screens the observed parameters after normalization. ≤ set the threshold, replace the original value of the observation parameter with the normalized value of the observation parameter, and update the original matrix of the observation parameter to a complete and ordered sequence matrix of the observation parameter. > Set the threshold, then adjust Distribution factor and The process factor repeats step S100;

[0035] (11);

[0036] The threshold is generally 0.15.

[0037] Furthermore, in step S200, a dual-gradient pressure-controlled drilling prediction method and a gas invasion condition overflow estimation method are established to form a drilling wellbore pressure dynamic control model;

[0038] S201, dual gradient managed pressure drilling prediction method is:

[0039] (1);

[0040] (2);

[0041] (3);

[0042] In formula (1), is the gas density in kg / m 3 ; is the acceleration due to gravity, m 2 / s; is the well inclination angle, unit (°); is the maximum drilling depth of a single pass, in m;

[0043] In formula (3), is the specific heat capacity, in J / (kg·℃); is the time within the calculation step, in seconds; is the axial length of the wellbore within the calculation step, in meters; is the thermal conductivity, in W / (m·K); is the radial length of the wellbore within the calculation step, in meters;

[0044] S202, the overflow estimation method for gas intrusion conditions is:

[0045] (4);

[0046] (5);

[0047] In formula (4), is the reservoir thickness within the calculation step, in m; is the reservoir permeability within the calculation step, in mD; is the gas phase viscosity, in mPa·s; is the gas hydrate compressibility factor, dimensionless; is the gas volume fraction, dimensionless;

[0048] In formula (5), is the wellbore volume, in m 3 .

[0049] The beneficial effects of step S200 are as follows: the dual gradient controlled pressure drilling prediction method is combined with the dual gradient drilling technology for modeling. The dual gradient drilling technology has broad development prospects in the field of natural gas hydrate drilling, can reduce the possibility of complex situations downhole, simplify the wellbore structure, and improve drilling efficiency. In addition, the dual gradient controlled pressure drilling prediction method covers Dual gradient drilling fluid variable liquid level height, The key parameter of the maximum drilling depth of a single pass can achieve the optimization design goal of accurately adjusting the variable liquid level of the dual-gradient drilling fluid and achieving the maximum drilling depth of a single pass; the overflow estimation method of the gas invasion condition comprehensively considers the secondary decomposition effect of the natural gas hydrate cuttings during the return process, and can identify the pseudo gas invasion and reservoir breathing effect during the drilling process; at the same time, the drilling wellbore pressure dynamic control model includes the parameters of the observation parameter sequence matrix and the state parameter sequence matrix, which fully considers LWD pressure data, Drilling fluid equivalent circulating density, Poor wellhead flow rate, The main factor of the additional density of cuttings can effectively improve the prediction accuracy of wellbore pressure in natural gas hydrate drilling.

[0050] Furthermore, step S300 is specifically as follows:

[0051] S301, establishing observation equation and state equation;

[0052] Observation equation: (14);

[0053] In formula (14), yes Observation parameters at time; yes State parameters at the moment; yes Observation error at time;

[0054] Equation of state: (15);

[0055] In formula (15), yes State parameters at the moment; yes The process error of the moment;

[0056] S302, using the particle filter algorithm, combined with the dual gradient pressure-controlled drilling prediction method, after the particle filter algorithm random sampling, updating, normalization, resampling a total of 4 steps, according to LWD pressure data, Drilling fluid equivalent circulating density, Drilling fluid mixed density, Wellhead back pressure, Ibukazu There are 6 observation parameters of cyclic friction to predict the next calculation step length. Predicted bottom hole pressure and Dual-gradient drilling fluid variable liquid level height; using particle filter algorithm, combined with gas invasion overflow estimation method, after random sampling, updating, normalization, resampling of particle filter algorithm, a total of 4 steps, according to Wellbore temperature, Annular pressure, Wellhead flow difference and The cuttings additional density is a total of 4 observation parameters to predict the next calculation step length Formation pressure and Gas intrusion rate;

[0057] S303, introducing statistical errors of observation parameters and state parameter statistical errors To evaluate the prediction effect, ≤ set threshold and ≤ the set threshold, the state parameter is output, and the threshold is generally 0.2; if the condition is not met, the resampling weight is adjusted and step S300 is repeated;

[0058] (16);

[0059] In formula (16), is the statistical error of the observed parameter, dimensionless; is the number of observed parameters; is the observed value obtained through the observation equation; is the value of the observation parameter sequence matrix; is the observation equation; is the standard value of error, dimensionless;

[0060] (17);

[0061] In formula (17), is the statistical error of the state parameter, dimensionless; is the expected value of the state parameter.

[0062] Furthermore, step S400 is specifically as follows:

[0063] S401, establish the restriction conditions of drilling safety density window and overflow safety evaluation standard;

[0064] (18);

[0065] Formula (18) is the drilling safety density window constraint condition. In formula (18), is the formation pore pressure, in MPa; is the formation collapse pressure, in MPa; is the formation fracture pressure, in MPa; is the formation loss pressure, in MPa;

[0066] (19);

[0067] Formula (19) is the overflow safety evaluation standard constraint condition. In formula (19), is the casing pressure, in MPa; is the riser pressure, in MPa; is the overflow drilling fluid density, unit: kg / m 3 ;

[0068] S402, according to the drilling safety density window restriction conditions and the overflow safety evaluation standard restriction conditions, verify the safety of the state parameters, if the conditions are met, output data, and update the original matrix of the state parameters to the state parameter sequence matrix; otherwise, adjust the state parameters in step S100. The threshold of the fitting error is set, and steps S100 to S400 are repeated.

[0069] The present invention also provides a computer program product, including a computer program / instruction, which, when executed by a processor, can implement the above-mentioned method for dynamically controlling wellbore pressure in natural gas hydrate drilling.

[0070] The beneficial effects of the present invention are embodied in:

[0071] (1) The present invention establishes the original matrix of observation parameters and the original matrix of state parameters, and proposes a better data normalization processing method based on adjacent wells and while-drilling data, which can solve the problem of lag and missing of observation parameters in natural gas hydrate drilling and production;

[0072] (2) The present invention establishes a dynamic control model for wellbore pressure in drilling that is suitable for dual-gradient drilling technology and gas invasion conditions, covers the main factors that affect the wellbore pressure in natural gas hydrate drilling, and can predict the wellbore pressure profile at any time during the natural gas hydrate drilling process;

[0073] (3) The present invention uses the particle filtering method to invert state parameters and conducts quantitative evaluation of the prediction effect, thereby improving the calculation accuracy of multiphase flow under complex constraints of natural gas hydrate drilling wellbore. Dual gradient drilling fluid variable liquid level height, Drilling fluid equivalent circulating density and Wellhead back pressure can increase the maximum drilling depth of a single drilling and improve drilling efficiency;

[0074] (4) The present invention adopts the dual constraints of safety density window and overflow to ensure that the optimized parameters can meet the drilling safety requirements. The calculation results can tend to be stable in a short time. The method has strong robustness and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The above and other features of the present invention will become more obvious through a detailed description of the embodiments shown in the accompanying drawings. The same reference numerals in the accompanying drawings of the present invention represent the same or similar elements. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative work. In the accompanying drawings:

[0076] Figure 1 A flow chart of a method for dynamically controlling wellbore pressure in natural gas hydrate drilling;

[0077] Figure 2 Schematic diagram of the calculation process of step S100 in Example 1;

[0078] Figure 3 Schematic diagram of the calculation process of step S300 in Example 1;

[0079] Figure 4 It is a verification curve diagram of bottom hole pressure and formation pressure in Example 1;

[0080] Figure 5 This is a comparison diagram for verifying the variable liquid level height and gas invasion rate of the dual-gradient drilling fluid in Example 1;

[0081] Figure 6 This is a design architecture diagram of a program product in Example 2. DETAILED DESCRIPTION

[0082] The following will be combined with the accompanying drawings and embodiments to clearly and completely describe the concept, specific technical means and technical effects of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0083] like Figure 1 The flowchart of a natural gas hydrate drilling wellbore pressure dynamic control method according to the present invention is shown below in conjunction with Figure 1 The specific technical means of a method for dynamically controlling wellbore pressure in natural gas hydrate drilling according to the present invention will be described.

[0084] The present invention provides a method for dynamically controlling the wellbore pressure of natural gas hydrate drilling, the method comprising the following steps:

[0085] S100, establishing an original matrix of observation parameters and an original matrix of state parameters, and performing data normalization processing of lag regression and missing complement on the observation parameters based on adjacent wells and while-drilling data, to form a pre-processed observation parameter sequence matrix;

[0086] S200, establishing a dual-gradient pressure-controlled drilling prediction method and a gas invasion condition overflow estimation method to form a drilling wellbore pressure dynamic control model;

[0087] S300, establishing observation equations and state equations, using particle filter algorithm, combined with the drilling wellbore pressure dynamic control model, predicting state parameters based on observation parameters, and evaluating the prediction effect of state parameters using statistical errors;

[0088] S400, verify the safety of the state parameters according to the drilling safety density window restriction and the overflow safety evaluation standard restriction, output data if the conditions are met, and update the original matrix of the state parameters to the state parameter sequence matrix.

[0089] Furthermore, step S100 is specifically as follows:

[0090] S101, establish the original matrix of observation parameters, with a total of 10 observation parameters, divided into dual-gradient pressure-controlled drilling observation parameters and gas invasion overflow observation parameters:

[0091] {y}_{e}=[\, {p}_{d}\, \, {\rho}_{ECD}\, \, \rho \, \, {p}_{w}\, \, H\, \, {p}_{f}\, ] (6);

[0092] {y}_{h}=[\, T\, \, {p}_{a}\, \, \Delta {Q}_{p}\, \, {\rho}_{a}\, ] (7);

[0093] In formula (6), It is the observation parameter of dual gradient managed pressure drilling; is the logging while drilling pressure data, unit: MPa; is the equivalent circulating density of drilling fluid, in kg / m 3 ; is the mixed density of drilling fluid, in kg / m 3 ; is the wellhead back pressure, in MPa; is the well depth, in m; is the cyclic friction, in MPa;

[0094] In formula (7), It is the overflow observation parameter of gas intrusion condition; is the wellbore temperature, in °C; is the annulus pressure, in MPa; is the wellhead flow difference, in L / s; is the additional density of cuttings, in kg / m 3 ;

[0095] S102, establish the original matrix of state parameters, with a total of 4 observation parameters, divided into dual-gradient pressure-controlled drilling state parameters and gas invasion overflow state parameters:

[0096] {x}_{e}=[\, {p}_{b}\, \, {H}_{x}\, ] (8);

[0097] {x}_{h}=[\, {p}_{s}\, \, {k}_{r}\, ] (9);

[0098] In formula (8), It is the state parameter of dual gradient managed pressure drilling; is the predicted bottom hole pressure, in MPa; is the variable liquid level height of dual gradient drilling fluid, in m;

[0099] In formula (9), It is the overflow state parameter of gas intrusion condition; is the formation pressure, in MPa; is the gas intrusion rate, in m 3 / s;

[0100] S103, based on the adjacent wells and while-drilling data, the observed parameters are subjected to data normalization processing of lag regression and missing data completion. Specifically, on the basis of considering the late arrival time factors of each observed parameter, the neighbor interpolation method and the complex-valued function fitting method are used to perform lag regression of the observed parameters, and to predict and fill in the missing data, so as to eliminate the dimensional influence between different attributes;

[0101] (10);

[0102] (12);

[0103] (13);

[0104] In formula (10), is the normalized value of the i-th observed parameter; It is the minimum value of the normalized value of the observed parameter within the calculation step; It is the maximum value of the normalized value of the observed parameter within the calculation step; It is the minimum value of the original value of the observed parameter within the calculation step; It is the maximum value of the original value of the observed parameter within the calculation step;

[0105] In formula (12), is the characteristic coefficient, dimensionless; is the total number of samples of the observed parameter, in units; is the original value of the ith observed parameter. When the observed parameter is missing, the data of the adjacent well at the same location is taken; is the average value of the original value of the observed parameter; is the average of the normalized values ​​of the observed parameters; is the distribution factor, the initial value is 1;

[0106] In formula (13), is the hysteresis coefficient, dimensionless; is the borehole diameter, in mm; is the drill diameter, in mm; is the pump displacement, in L / s; is the mechanical drilling speed, in m / s; is the distance between the sensor and the calculation point, in meters; is the process factor, and its initial value is 1;

[0107] S104, introduced The fitting error is The fitting error screens the observed parameters after normalization. ≤ set the threshold, replace the original value of the observation parameter with the normalized value of the observation parameter, and update the original matrix of the observation parameter to a complete and ordered sequence matrix of the observation parameter. > Set the threshold, then adjust Distribution factor and The process factor repeats step S100;

[0108] (11);

[0109] The threshold is generally 0.15.

[0110] Furthermore, in step S200, a dual-gradient pressure-controlled drilling prediction method and a gas invasion condition overflow estimation method are established to form a drilling wellbore pressure dynamic control model;

[0111] S201, dual gradient managed pressure drilling prediction method is:

[0112] (1);

[0113] (2);

[0114] (3);

[0115] In formula (1), is the gas density in kg / m 3 ; is the acceleration due to gravity, m 2 / s; is the well inclination angle, unit (°); is the maximum drilling depth of a single pass, in m;

[0116] In formula (3), is the specific heat capacity, in J / (kg·℃); is the time within the calculation step, in seconds; is the axial length of the wellbore within the calculation step, in meters; is the thermal conductivity, in W / (m·K); is the radial length of the wellbore within the calculation step, in meters;

[0117] S202, the overflow estimation method for gas intrusion conditions is:

[0118] (4);

[0119] (5);

[0120] In formula (4), is the reservoir thickness within the calculation step, in m; is the reservoir permeability within the calculation step, in mD; is the gas phase viscosity, in mPa·s; is the gas hydrate compressibility factor, dimensionless; is the gas volume fraction, dimensionless;

[0121] In formula (5), is the wellbore volume, in m 3 .

[0122] Furthermore, step S300 is specifically as follows:

[0123] S301, establishing observation equation and state equation;

[0124] Observation equation: (14);

[0125] In formula (14), yes Observation parameters at time; yes State parameters at the moment; yes Observation error at time;

[0126] Equation of state: (15);

[0127] In formula (15), yes State parameters at the moment; yes The process error of the moment;

[0128] S302, using the particle filter algorithm, combined with the dual gradient pressure-controlled drilling prediction method, after the particle filter algorithm random sampling, updating, normalization, resampling a total of 4 steps, according to LWD pressure data, Drilling fluid equivalent circulating density, Drilling fluid mixed density, Wellhead back pressure, Ibukazu There are 6 observation parameters of cyclic friction to predict the next calculation step length. Predicted bottom hole pressure and Dual-gradient drilling fluid variable liquid level height; using particle filter algorithm, combined with gas invasion overflow estimation method, after random sampling, updating, normalization, resampling of particle filter algorithm, a total of 4 steps, according to Wellbore temperature, Annular pressure, Wellhead flow difference and The cuttings additional density is a total of 4 observation parameters to predict the next calculation step length Formation pressure and Gas intrusion rate;

[0129] S303, introducing statistical errors of observation parameters and state parameter statistical errors To evaluate the prediction effect, ≤ set threshold and ≤ the set threshold, the state parameter is output, and the threshold is generally 0.2; if the condition is not met, the resampling weight is adjusted and step S300 is repeated;

[0130] (16);

[0131] In formula (16), is the statistical error of the observed parameter, dimensionless; is the number of observed parameters; is the observed value obtained through the observation equation; is the value of the observation parameter sequence matrix; is the observation equation; is the standard value of error, dimensionless;

[0132] (17);

[0133] In formula (17), is the statistical error of the state parameter, dimensionless; is the expected value of the state parameter.

[0134] Furthermore, step S400 is specifically as follows:

[0135] S401, establish the restriction conditions of drilling safety density window and overflow safety evaluation standard;

[0136] (18);

[0137] Formula (18) is the drilling safety density window constraint condition. In formula (18), is the formation pore pressure, in MPa; is the formation collapse pressure, in MPa; is the formation fracture pressure, in MPa; is the formation loss pressure, in MPa;

[0138] (19);

[0139] Formula (19) is the overflow safety evaluation standard constraint condition. In formula (19), is the casing pressure, in MPa; is the riser pressure, in MPa; is the overflow drilling fluid density, unit: kg / m 3 ;

[0140] S402, according to the drilling safety density window restriction conditions and the overflow safety evaluation standard restriction conditions, verify the safety of the state parameters, if the conditions are met, output data, and update the original matrix of the state parameters to the state parameter sequence matrix; otherwise, adjust the state parameters in step S100. The threshold of the fitting error is set, and steps S100 to S400 are repeated.

[0141] Example 1

[0142] This embodiment discloses a method for dynamically controlling the wellbore pressure of a natural gas hydrate drilling. The basic information of the natural gas hydrate reservoir in this embodiment is as follows:

[0143] A natural gas hydrate pilot production area in the Shenhu sea area on the northern slope of the Baiyun Sag in the Pearl River Mouth Basin in the northern South China Sea has a water depth of 1,000 to 1,300 meters. Natural gas hydrates are mainly found in unconsolidated sediments of Quaternary mud silt or silt mud, and are produced 207 to 297 meters below the mud line. The data of adjacent wells are complete.

[0144] The natural gas hydrate reservoir parameters are shown in Table 1:

[0145] Table 1 Natural gas hydrate reservoir parameters

[0146]

[0147] Porosity and permeability are average values.

[0148] The test well body structure is shown in Table 2:

[0149] Table 2 Wellbore structure of the test well

[0150]

[0151] The casing depth is the length of the wellbore below the mudline.

[0152] Furthermore, the reservoir 207 to 297 meters below the mudline is taken as the research object. The horizon is the open hole section of the third opening. The drilling data from 247 to 273 meters below the mudline is missing. The calculation step is 5 meters, so the 10 observation parameters are all 18-dimensional vectors. For example, {p}_{d}={[\, {p}_{d1},\, \, {p}_{d2},\, {p}_{d3},\, \cdot \cdot \cdot ,\, {p}_{d18}\, ]}^{T} ,in, to For data missing segments, assign null values. to , to Fill with original value.

[0153] Furthermore, the original matrix of observation parameters is established, specifically:

[0154] {y}_{e}=[\, {p}_{d}\, \, {\rho}_{ECD}\, \, \rho \, \, {p}_{w}\, \, H\, \, {p}_{f}\, ] ;

[0155] {y}_{h}=[\, T\, \, {p}_{a}\, \, \Delta {Q}_{p}\, \, {\rho}_{a}\, ] ;

[0156] in, is an 18×6 matrix, It is an 18×4 matrix.

[0157] Furthermore, with a calculation step length of 5 meters, the four observation parameters are all 18-dimensional vectors, all data are assigned null values, and the original matrix of state parameters is established, specifically:

[0158] {x}_{e}=[\, {p}_{b}\, \, {H}_{x}\, ] ;

[0159] {x}_{h}=[\, {p}_{s}\, \, {k}_{r}\, ] ;

[0160] in, and Both are 18×2 matrices.

[0161] Furthermore, the hysteresis is 5 meters, the average drilling speed is 21 meters per hour, and the formulas (10) (11) (12) (13) of the present invention are used. According to Figure 2 The calculation process shown in the figure uses a 5-meter calculation step to perform data normalization processing on the observed parameters by performing lag regression and missing completion, and obtains The maximum value is 0.11, which updates the original matrix of observation parameters to a complete and ordered sequence matrix of observation parameters.

[0162] Furthermore, the dynamic control model of drilling wellbore pressure is established by using the formulas (1), (2), (3), (4), and (5) of the present invention. The basic data used are shown in Tables 3 and 4:

[0163] Table 3 Basic data used to establish the dual gradient managed pressure drilling prediction method

[0164] <![CDATA[ ρ g Gas density (kg / m 3 ） ]]> <![CDATA[ C p Specific heat capacity (J / (kg·℃)) ]]> <![CDATA[ 1.54×10 3 ]]> <![CDATA[ H x Dual gradient drilling fluid variable liquid level height (m) ]]> <![CDATA[ H max Maximum drilling depth per single pass (m) ]]> <![CDATA[ 3.93×10 4 ]]>

[0165] Table 3 and The data are the initial values ​​of the boundary conditions;

[0166] Table 4 Basic data used to establish the overflow estimation method for gas intrusion conditions

[0167] <![CDATA[ m 0 Reservoir thickness within calculation step (m) ]]> <![CDATA[ u g Gas phase viscosity (mPa·s) ]]> <![CDATA[ 3.93×10 -3 ]]> <![CDATA[ α g Gas volume fraction ]]> <![CDATA[ V Wellbore volume (m 3 ） ]]>

[0168] Table 4 It is the value under standard condition.

[0169] Further, according to Figure 3 The calculation process shown in the figure adopts formula (14) (15) of the present invention, takes 5 meters as the calculation step, uses the observed parameters of the current calculation step to invert the state parameters of the next calculation step, and adopts formula (16) (17) of the present invention to calculate Observation parameter statistical errors and The statistical error of state parameters is obtained =0.13, =0.09, because and If both are less than 0.2, the state parameters are calculated.

[0170] Furthermore, the formulas (18) and (19) of the present invention are used to verify that the state parameters meet the drilling safety density window and overflow safety evaluation constraints, then the state parameter sequence matrix is ​​output and updated. The calculation results are Figure 4 and Figure 5 Displayed in.

[0171] like Figure 4 As shown, the predicted bottom hole pressure and formation pressure obtained by the method of the present invention are complete, with a calculation step of 5 meters and a total of 28 data. The similarity between the distribution characteristics of the predicted bottom hole pressure and the changing trend of the measured bottom hole pressure is 83%, and the predicted formation pressure is close to the lower limit of the safety density window, indicating that the method of the present invention solves the problems of lag and missing observation parameters, and the established dynamic control model of drilling wellbore pressure can accurately invert the overall trend of natural gas hydrate drilling wellbore pressure. The adopted particle filtering algorithm improves the calculation accuracy and can predict the wellbore pressure profile at any time during the drilling process.

[0172] like Figure 4As shown, in the natural gas hydrate reservoir section from 278 to 297 meters, the measured value of the wellbore pressure fluctuates and is severely disturbed. The method of the present invention can solve it well, and the calculation results converge quickly in a short time.

[0173] like Figure 4 As shown in the figure, the minimum predicted bottom hole pressure is 63MPa and the maximum is 79MPa. The predicted bottom hole pressure values ​​are all within the safety density window, and the predicted formation pressure values ​​are slightly greater than the lower limit of the safety density window, ensuring that the optimized parameters meet the drilling safety requirements and can safely guide production applications.

[0174] like Figure 5 As shown in the figure, the change of the liquid level height and gas invasion rate of the dual-gradient drilling fluid with the well depth is demonstrated. With the increase of the drilled reservoir thickness and the secondary decomposition of the cuttings during the return process, a sudden gas invasion occurs in the gaseous hydrocarbon layer at 255 to 270 meters, which is in line with the actual situation. The calculated maximum gas invasion rate is 0.4m / s.

[0175] like Figure 5 As shown, in response to gas invasion and wellbore pressure changes, the method of the present invention adjusts the dual-gradient drilling fluid variable liquid level height 6 times in total, with an adjustment range of 53 to 60 meters, and the adjustment starting point is earlier than the gas invasion rate change point, thereby improving the well control capability.

[0176] In Example 1, by optimizing Dual gradient drilling fluid variable liquid level height, Drilling fluid equivalent circulating density and Based on the wellhead back pressure, it is calculated that the maximum drilling depth for a single drilling is 365 meters, which can achieve one-trip drilling of the target layer, effectively shorten the drilling cycle, reduce drilling costs, and reduce underground complexities.

[0177] Example 2

[0178] This embodiment discloses a natural gas hydrate drilling wellbore pressure dynamic control program product. The design architecture of the program product is as follows: Figure 6 shown.

[0179] The data standard processing unit is as follows: establishing the original matrix of observation parameters and the original matrix of state parameters, and performing data standardization processing of lag regression and missing completion on the observation parameters based on the adjacent well and drilling data to form a preprocessed observation parameter sequence matrix;

[0180] Furthermore, some code examples of Java implementation for lagged regression and missing completion of observed parameters can be provided:

[0181] public void lagRegression( ) {

[0182] divideCurvatureA = new double[n + 1];

[0183] for (int i = 1; i <a.length; i++) {divideCurvatureA[i] = (a[i]- a[i - 1]) / (s[i]- s[i - 1]);}

[0184] for (int i = 1; i <s.length; i++) {

[0185] divideCurvature[i] = Math.sqrt(divideCurvatureA[i]*divideCurvatureA[i]

[0186] + divideCurvatureB[i]*divideCurvatureB[i]*Math.sin(a[i])

[0187] *Math.sin(a[i]));}}

[0188] Among them, the lagRegression function can traverse the original matrix of observation parameters, call the neighboring interpolation function and the complex value function, perform lag regression on the observation parameters in turn, predict and fill in the missing data, and compare The fitting error filters the results to achieve the established calculation function.

[0189] Pressure control prediction model unit: establish a dual-gradient pressure control drilling prediction method and a gas invasion condition overflow estimation method to form a drilling wellbore pressure dynamic control model.

[0190] Particle filter inversion unit: establish observation equations and state equations, use particle filter algorithm, combine drilling wellbore pressure dynamic control model, predict state parameters based on observation parameters, and use statistical errors to evaluate the prediction effect of state parameters;

[0191] Furthermore, some code examples of Java implementation of using particle filtering to predict state parameters based on observed parameters can be provided:

[0192] public boolean analysisStress(double l) {

[0193] for (int i = 0; i <hl_in.length; i++) {

[0194] hl_in[i] = Pi0*(r_out+r_in) / (r_out-r_in) - 2*r_out*Po0 / (r_out-r_in);

[0195] hl_out[i] = Po0*(r_out+r_in) / (r_out-r_in) - 2*r_in*Pi0 / (r_out-r_in);

[0196] optimizeZl[i] = - Double.parseDouble(df.format(zl[i] / Math.pow(10, 2)));

[0197] optimizeWl_in[i] = Double.parseDouble(df.format(wl_in[i] / Math.pow(10, 12)));

[0198] yl_1 = Math.sqrt((Math.pow((optimizeZl[i] + optimizeWl_in[i]-jl_in[i]), 2)

[0199] + Math.pow((optimizeZl[i]+ optimizeWl_in[i]- hl_in[i]),2)

[0200] + Math.pow((jl_in[i]- hl_in[i]), 2)) / 2);

[0201] yl_2 = Math.sqrt((Math.pow((optimizeZl[i] - optimizeWl_in[i]-jl_in[i]), 2)

[0202] + Math.pow((optimizeZl[i]- optimizeWl_in[i]- hl_in[i]), 2)

[0203] + Math.pow((jl_in[i]- hl_in[i]), 2)) / 2);}}

[0204] The analysisStress function describes the probability density function through a random sample set, and uses the mean of the samples instead of the integral operation. As the observation parameters are updated, the weights and positions of the particles are continuously adjusted, and the prior distribution is corrected and updated, and finally the posterior probability distribution of the state parameters is obtained.

[0205] Safety verification output unit: Verify the safety of state parameters according to the drilling safety density window restriction conditions and overflow safety evaluation standard restriction conditions. If the conditions are met, output data and update the original state parameter matrix to the state parameter sequence matrix.

[0206] It will be apparent that in some embodiments, computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0207] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that a technician familiar with the technical field can think of within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for dynamically controlling wellbore pressure in natural gas hydrate drilling, characterized in that: The following steps are involved: S100, establishing an original matrix of observation parameters and an original matrix of state parameters, and performing data normalization processing of lag regression and missing complement on the observation parameters based on adjacent wells and while-drilling data, to form a pre-processed observation parameter sequence matrix; S200, establishing a dual-gradient pressure-controlled drilling prediction method and a gas invasion condition overflow estimation method to form a drilling wellbore pressure dynamic control model; Among them, the dual-gradient managed pressure drilling prediction method is specifically as follows: （1）； （2）； （3）； In formula (1), is the predicted bottom hole pressure, in MPa; is the logging while drilling pressure data, unit: MPa; is the gas density in kg / m 3 ; is the acceleration due to gravity, m 2 / s; is the variable liquid level height of dual gradient drilling fluid, in m; is the well inclination angle, unit (°); is the drilling fluid mixed density, unit: kg / m 3 ; is the well depth, in m; is the wellhead back pressure, in MPa; is the maximum drilling depth of a single pass, in m; is the cyclic friction, in MPa; In formula (2), is the equivalent circulating density of drilling fluid, in kg / m 3 ; In formula (3), is the specific heat capacity, in J / (kg·℃); is the wellbore temperature, in °C; is the time within the calculation step, in seconds; is the axial length of the wellbore within the calculation step, in meters; is the thermal conductivity, in W / (m·K); is the radial length of the wellbore within the calculation step, in meters; The specific method for estimating overflow under gas intrusion conditions is as follows: （4）； （5）； In formula (4), is the gas intrusion rate, in m 3 / s; is the reservoir thickness within the calculation step, in m; is the wellhead flow difference, in L / s; is the annulus pressure, in MPa; is the additional density of cuttings, in kg / m 3 ; is the reservoir permeability within the calculation step, in mD; is the gas phase viscosity, in mPa·s; is the gas hydrate compressibility factor, dimensionless; is the borehole diameter, in mm; is the formation pressure, in MPa; is the gas volume fraction, dimensionless; In formula (5), is the wellbore volume, in m 3 ; S300, establishing observation equations and state equations, using particle filter algorithm, combined with the drilling wellbore pressure dynamic control model, predicting state parameters based on observation parameters, and evaluating the prediction effect of state parameters using statistical errors; S400, verify the safety of the state parameters according to the drilling safety density window restriction and the overflow safety evaluation standard restriction, output data if the conditions are met, and update the original matrix of the state parameters to the state parameter sequence matrix.

2. A natural gas hydrate drilling wellbore pressure dynamic control method according to claim 1, characterized in that: The step S100 includes: S101, establish the original matrix of observation parameters: （6）； （7）； In formula (6), It is the observation parameter of dual gradient managed pressure drilling; In formula (7), It is the overflow observation parameter of gas intrusion condition; S102, establish the original matrix of state parameters: （8）； （9）； In formula (8), It is the state parameter of dual gradient managed pressure drilling; In formula (9), It is the overflow state parameter of gas intrusion condition; S103, based on the data from adjacent wells and while-drilling, the observed parameters are subjected to data normalization processing by means of lag regression and missing complementation. The calculation method is: （10）； In formula (10), is the normalized value of the i-th observed parameter; It is the minimum value of the normalized value of the observed parameter within the calculation step; It is the maximum value of the normalized value of the observed parameter within the calculation step; It is the maximum value of the original value of the observed parameter within the calculation step; It is the minimum value of the original value of the observed parameter within the calculation step; is the characteristic coefficient, dimensionless; is the original value of the ith observed parameter. When the observed parameter is missing, the data of the adjacent well at the same location is taken; is the hysteresis coefficient, dimensionless; S104, introduced The fitting error is The fitting error screens the observed parameters after normalization. ≤ set the threshold, replace the original value of the observation parameter with the normalized value of the observation parameter, and update the original matrix of the observation parameter to a complete and ordered sequence matrix of the observation parameter. > Set the threshold, then adjust Distribution factor and The process factor repeats step S100; （11）； In formula (11), is the total number of samples of the observed parameter, in units; The threshold is generally 0.

15.

3. A natural gas hydrate drilling wellbore pressure dynamic control method according to claim 2, characterized in that: In step S103, the characteristic coefficient The calculation method is: （12）； In formula (12), is the average value of the original value of the observed parameter; is the average of the normalized values ​​of the observed parameters; is the distribution factor, and its initial value is 1.

4. A natural gas hydrate drilling wellbore pressure dynamic control method according to claim 2, characterized in that: In step S103, the hysteresis coefficient The calculation method is: （13）； In formula (13), is the drill diameter, in mm; is the pump displacement, in L / s; is the process factor, and its initial value is 1; is the mechanical drilling speed, in m / s; is the distance between the sensor and the calculation point, in meters.

5. A natural gas hydrate drilling wellbore pressure dynamic control method according to claim 1, characterized in that: The step S300 includes: S301, establishing observation equation and state equation; Observation equation: (14); In formula (14), yes Observation parameters at time; yes State parameters at the moment; yes Observation error at time; Equation of state: (15); In formula (15), yes State parameters at the moment; yes The process error of the moment; S302, using the particle filter algorithm, combined with the dual gradient pressure-controlled drilling prediction method, after the particle filter algorithm random sampling, updating, normalization, resampling a total of 4 steps, according to LWD pressure data, Drilling fluid equivalent circulating density, Drilling fluid mixed density, Wellhead back pressure, Ibukazu There are 6 observation parameters of cyclic friction to predict the next calculation step length. Predicted bottom hole pressure and Dual gradient drilling fluid variable liquid level height; using particle filter algorithm, combined with gas invasion overflow estimation method, after random sampling, updating, normalization, resampling of particle filter algorithm, a total of 4 steps, according to Wellbore temperature, Annular pressure, Wellhead flow difference and The cuttings additional density is a total of 4 observation parameters to predict the next calculation step length Formation pressure and Gas intrusion rate; S303, introducing statistical errors of observation parameters and state parameter statistical errors To evaluate the prediction effect, ≤ set threshold and ≤ the set threshold, the state parameter is output, and the threshold is generally 0.2; if the condition is not met, the resampling weight is adjusted and step S300 is repeated; （16）； In formula (16), is the statistical error of the observed parameter, dimensionless; is the number of observed parameters; is the observed value obtained through the observation equation; is the value of the observation parameter sequence matrix; is the observation equation; is the standard value of error, dimensionless; （17）； In formula (17), is the statistical error of the state parameter, dimensionless; is the expected value of the state parameter.

6. A natural gas hydrate drilling wellbore pressure dynamic control method according to claim 1, characterized in that: In step S400, the overflow safety evaluation standard restriction conditions are: （19）； In formula (19), is the casing pressure, in MPa; is the riser pressure, in MPa; is the overflow drilling fluid density, in kg / m 3 .

7. A computer program product comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by a processor, the method steps described in any one of claims 1 to 6 are implemented.

Images